Apakah AI Akan Menggantikan Pekerjaan Data Analyst di Tahun 2026

Scroll LinkedIn sebentar aja, pasti nemu minimal 5 post yang bikin panik: "AI bakal gantikan semua jobs!", "Data analyst tinggal tunggu waktu!", atau yang paling dramatis: "ChatGPT bisa analyze data dalam 10 detik, why hire humans?"

Gue ngerti kok, kekhawatiran lo valid banget. Apalagi kalau lo lagi career switch ke data analyst atau baru mulai belajar, terus liat AI bisa bikin visualisasi data, generate insights, bahkan nulis SQL queries. Rasanya kayak, "Waduh, gue belum mulai udah mau digantiin nih?"

Tapi tunggu dulu.

Bagian 1: Kenapa Semua Orang Panic Soal AI?

Fear vs Reality

Gue pernah liat demo ChatGPT analyze dataset sales, generate insights, bikin visualization, complete dalam 30 detik. Impressive? Absolutely. Bikin takut? Maybe. Tapi ini yang perlu lo tau: impressive demo ≠ real-world application.

Di real job, data analyst bukan cuma "analyze data terus done". Ada banyak layer yang ga keliatan di demo 30 detik itu:

Yang Keliatan di Demo AI:

  • Input clean dataset
  • Get instant insights
  • Beautiful charts
  • Done!

Reality di Kantor:

  • Data berantakan dari 5 sistem berbeda
  • Half data missing, half data ga masuk akal
  • Stakeholder ga jelas mau apa
  • Requirements berubah tengah jalan
  • Harus explain findings ke bos yang ga ngerti technical
  • Politics: tim A dan tim B debat soal definisi "active user"

See the difference? AI bisa bantu dengan technical tasks, tapi ga bisa navigate complexity real-world business environment.

Apa Sih Sebenarnya yang Dikerjakan Data Analyst?

Ini misunderstanding terbesar. Banyak yang mikir data analyst tuh cuma:

  1. Open Excel/SQL
  2. Run queries
  3. Bikin chart
  4. Send report

Kalau cuma itu, ya emang AI bisa gantiin. But that's not the full picture.

Real Day in the Life:

Morning (09:00 - 12:00):

  • Meeting dengan product team: mereka mau tau kenapa conversion rate drop 15%
  • Bukan cuma kasih angka, tapi investigate: apakah karena UI change? Seasonal? Bug? Marketing campaign berubah?
  • Dig deeper ke data, cross-check sama customer support tickets
  • Hipotesis: kayaknya ada issue di mobile checkout

Afternoon (13:00 - 17:00):

  • Validate hipotesis dengan segmented analysis
  • Ternyata bener: mobile conversion drop 40%, desktop aman
  • Collaborate sama engineering: ada bug di payment gateway mobile
  • Prepare presentation buat leadership: explain root cause, impact, dan recommended actions
  • Kasih business context: "We're losing $50k daily, need immediate fix"

Yang AI Bisa:

  • Generate initial metrics (conversion rate changes)
  • Create standard visualizations
  • Basic pattern detection

Yang AI Gabisa:

  • Tau bahwa perlu cross-check dengan support tickets
  • Understand bahwa mobile dan desktop perlu dipisah
  • Communicate urgency ke leadership dengan proper context
  • Collaborate dengan engineering buat troubleshoot
  • Know what actions are realistic given company resources

Value Sebenarnya Data Analyst

Lo bukan cuma "orang yang bikin chart". Lo adalah translator antara data dengan business decisions.

Technical Skills (30% of the job):

  • SQL, Python, Excel
  • Data cleaning & processing
  • Statistical analysis
  • Visualization

Business Skills (40% of the job):

  • Understand company metrics yang penting
  • Know business context
  • Identify actionable insights (bukan cuma "interesting" insights)
  • Prioritize analysis based on impact

Communication Skills (30% of the job):

  • Translate technical findings ke bahasa manusia
  • Storytelling dengan data
  • Create compelling presentations
  • Handle stakeholder dengan expectations berbeda-beda

AI bisa bantu dengan yang 30% pertama. Tapi 70% sisanya? That's all human.

Kenapa Topik Ini Penting Banget

Job Market Reality Check:

Kalau AI beneran mau gantikan data analyst, harusnya hiring udah drop drastis kan?

Faktanya:

  • LinkedIn data (2025): Data Analyst job postings naik 35% dari tahun lalu
  • Salary range Indonesia: Rp 8-25 juta/bulan untuk mid-level
  • Glassdoor: Data Analyst masuk "Top 15 Best Jobs in Tech 2025"

Proyeksi Pertumbuhan:

  • U.S. Bureau of Labor Statistics: 23% growth hingga 2031
  • McKinsey: Demand untuk "analytics-skilled workers" terus naik
  • Deloitte: Data analytics jadi top 3 skill yang dicari companies

Yang berubah bukan demand-nya, tapi skill requirements-nya. Companies sekarang cari data analyst yang:

  • Bisa pake AI tools efficiently
  • Punya strong business acumen
  • Excel di communication
  • Adaptable ke new technologies

Set Ekspektasi Artikel Ini

Gue ga akan bullshit atau kasih false hope. Artikel ini based on:

Research & Data:

  • McKinsey Global Institute studies
  • World Economic Forum reports
  • Gartner predictions
  • Real job market data

Real Examples:

  • Case studies dari tech companies
  • Examples tools yang actually useful
  • Realistic skill development path

Actionable Advice:

  • Konkret steps untuk upgrade skills
  • Free resources (including BuildWithAngga classes)
  • Timeline yang reasonable

Let's dive in dengan data, bukan asumsi.

Bagian 2: Riset Bilang - AI Enggak Akan Gantikan Data Analyst Sepenuhnya

Okay, lupakan opini random di internet. Mari kita liat apa kata research institutions dan tech leaders yang actually study this shit for a living.

McKinsey Global Institute: The Automation Reality

McKinsey nge-release comprehensive study tentang automation dan future of work. Key findings yang relevan:

"Only 5% of occupations can be fully automated with current technology"

Mayoritas pekerjaan itu kombinasi tasks yang bisa dan ga bisa di-automate. Data analyst masuk kategori "augmented jobs" - pekerjaan yang di-enhance oleh AI, bukan digantikan.

Breakdown Tasks Data Analyst:

  • 40% technical tasks (bisa di-automate/augment)
  • 35% judgment & decision making (ga bisa di-automate)
  • 25% stakeholder management (ga bisa di-automate)

Artinya? Best case scenario, AI bisa handle 40% tasks. Sisanya tetep butuh human.

Quote from McKinsey Research:

"The future of work is not about humans vs. machines, but humans + machines. Jobs that combine human judgment with AI capabilities will see the most growth."

Gartner Report: AI-Assisted, Not AI-Replaced

Gartner, perusahaan research yang literally jobnya predict tech trends, punya data menarik:

Prediksi 2026:

  • 80% of data analytics tasks akan AI-assisted
  • 90% of organizations akan require human oversight untuk AI-generated insights
  • Zero major organizations akan fully automate data analysis roles

Kenapa Human Oversight Mandatory?

Gartner identify critical risks kalau pure AI tanpa human:

  • Bias amplification: AI bisa amplify existing bias di data
  • Context misinterpretation: AI ga paham business nuance
  • Accountability issues: Kalau AI salah, who's responsible?
  • Ethical concerns: Some decisions butuh moral judgment

Real Example dari Gartner Case Study:

E-commerce company pake AI buat predict customer churn. AI recommend: "Target customers dengan discount 50%". Sounds good?

Human analyst noticed:

  • AI include loyal customers yang ga butuh discount
  • Budget blow up 300%
  • Margin jadi negative

Human intervention saved company millions. AI kasih recommendation, human decide if it makes sense.

World Economic Forum: Future of Jobs Report 2025

WEF nge-track employment trends globally. Data Analyst specifically disebutin sebagai "Growing Role" bukan "Declining Role".

Expected Growth by 2027:

  • Data Analysts & Scientists: +31%
  • AI & Machine Learning Specialists: +40%
  • Business Intelligence Analysts: +25%

Notice something? Semua data-related roles tumbuh. Kenapa? Because data volume explodes faster than AI can fully handle it.

Changing Skill Requirements:

2020 Top Skills2026 Required Skills
ExcelExcel + AI tools
SQLSQL + Python
TableauTableau + ML basics
StatisticsStatistics + AI literacy
-Prompt engineering
-AI output validation

It's not replacement, it's skill evolution.

Real Examples dari Tech Companies

Google:

Google punya internal tools kayak "Gemini for Workspace" yang bantu data analysts. Tapi mereka ga reduce headcount. Instead:

  • Analysts bisa handle 3x more projects
  • Focus shift dari "data processing" ke "strategic insights"
  • New roles created: "AI-assisted Analytics Lead"

Netflix:

Netflix terkenal data-driven. They use AI extensively buat recommendations. But:

  • Data analyst team actually growing
  • AI generate baseline insights
  • Humans refine, validate, dan translate to product decisions
  • Quote dari VP Analytics: "AI is our junior analyst that works 24/7. But we still need senior analysts to make sense of it all."

Tokopedia (Local Example):

Tokopedia pake AI tools buat automated reporting. Impact:

  • Report generation time: dari 2 hari jadi 2 jam
  • Analyst freed up untuk deeper analysis
  • Result: Better product decisions, faster iteration
  • Headcount: Tetap, bahkan nambah di beberapa teams

Pattern Across Companies:

  1. AI adoption increases analyst productivity
  2. Time saved digunakan untuk higher-value work
  3. New roles emerge (AI-assisted analyst, ML analyst, etc.)
  4. Total employment di data teams stable atau naik

Interview Insights dari Industry

Gue collect beberapa perspectives dari actual data analysts:

Sarah, Senior Data Analyst di Tech Startup:

"AI tools kayak ChatGPT Code Interpreter bantu banget buat quick exploratory analysis. Tapi untuk strategic decisions? Masih butuh manusia yang paham business context. AI bisa bilang 'sales turun 20%', tapi ga bisa bilang 'oh ini karena competitor baru launch promo'."

Rizki, Data Analyst di E-commerce:

"Dulu gue spend 60% waktu buat data cleaning dan basic processing. Sekarang dengan AI tools, jadi 20%. Sisa 80% gue pake buat deep analysis dan present findings ke stakeholders. Kerjaan gue ga ilang, malah jadi lebih interesting."

Linda, Analytics Manager:

"Team gue sekarang hire analysts yang bisa leverage AI tools. Bukan yang takut AI. Yang adaptable, mau belajar tools baru, dan bisa combine AI outputs dengan critical thinking - those are the winners."

The Data Speaks

Semua research consensus ke satu point:

AI Will Transform, Not Replace

  • Transform: How analysts work, tools they use, speed of analysis
  • Not Replace: The need for human judgment, business acumen, communication

Lo masih takut? Let me put it this way:

Saat Excel introduced, orang bilang accountant bakal hilang. What happened?

  • Accountants masih ada
  • Mereka jadi lebih produktif
  • Focus shift dari manual calculation ke analysis & advisory

Same thing dengan AI untuk data analysts.

Bagian 3: AI Bikin Data Analyst Kerja Lebih Cepat dan Presisi

Okay, jadi AI ga akan gantikan kerjaan lo. Tapi gimana sih AI actually membantu? Let me show you real productivity gains.

Before AI vs With AI: Real Task Breakdown

Scenario: E-commerce Monthly Performance Analysis

BEFORE AI (Traditional Way):

Week 1 - Data Collection & Cleaning:

  • Day 1-2: Extract data dari berbagai sources (database, Google Analytics, CRM)
  • Manual query writing: 4 jam
  • Data cleaning: identify missing values, outliers, inconsistencies: 6 jam
  • Data type conversions dan formatting: 2 jam
  • Total: 12 jam kerja

Week 2 - Exploratory Analysis:

  • Day 3: Calculate basic metrics (conversion rate, AOV, retention): 3 jam
  • Day 4: Segment analysis (by region, device, user type): 4 jam
  • Day 5: Create visualizations manually: 4 jam
  • Identify trends dan patterns: 3 jam
  • Total: 14 jam kerja

Week 3 - Deep Dive & Reporting:

  • Day 6-7: Investigate anomalies: 6 jam
  • Statistical testing: 4 jam
  • Create presentation slides: 5 jam
  • Write narrative dan recommendations: 3 jam
  • Total: 18 jam kerja

GRAND TOTAL: 44 jam (lebih dari 1 minggu full-time work)

WITH AI TOOLS (2026 Way):

Day 1 - Data Preparation:

  • Upload data ke AI platform
  • AI-automated profiling: 5 menit (identify missing values, outliers, data types)
  • Review AI suggestions untuk cleaning: 15 menit
  • One-click apply fixes: 5 menit
  • Manual validation: 30 menit
  • Total: ~1 jam

Day 2 - AI-Assisted Analysis:

  • Prompt AI: "Generate exploratory analysis for e-commerce performance"
  • AI generates: basic metrics, segments, visualizations: 10 menit
  • Review dan validate outputs: 30 menit
  • Deep dive ke interesting patterns AI found: 2 jam
  • Ask follow-up questions ke AI: 30 menit
  • Total: ~3 jam

Day 3 - Insights & Reporting:

  • AI generate first draft of report: 5 menit
  • Refine insights dengan business context: 1 jam
  • Customize visualizations: 30 menit
  • Add strategic recommendations (human part): 1 jam
  • Final review: 30 menit
  • Total: ~3 jam

GRAND TOTAL: 7 jam (kurang dari 1 hari kerja)

Productivity Gain: 84% time saved!

Real Task Automation Examples

1. Data Cleaning (Konkret Example)

Traditional Method:

# Manual detection dan fixing
import pandas as pd

df = pd.read_csv('sales_data.csv')

# Manual check missing values
print(df.isnull().sum())

# Manual decide how to handle each column
df['price'].fillna(df['price'].mean(), inplace=True)
df['category'].fillna('Unknown', inplace=True)

# Manual outlier detection
Q1 = df['sales'].quantile(0.25)
Q3 = df['sales'].quantile(0.75)
IQR = Q3 - Q1
outliers = df[(df['sales'] < Q1 - 1.5*IQR) | (df['sales'] > Q3 + 1.5*IQR)]

# Dan seterusnya... hundreds of lines

With AI Tools:

# Using AI-powered tools like pandas-ai
import pandas as pd
from pandasai import SmartDataframe

df = pd.read_csv('sales_data.csv')
sdf = SmartDataframe(df)

# Natural language query
sdf.chat("Clean this dataset, handle missing values appropriately, and remove outliers")

# AI automatically:
# - Detects data types
# - Suggests appropriate imputation methods
# - Identifies and handles outliers
# - Generates cleaning report

Time saved: dari 3 jam → 15 menit

2. Exploratory Data Analysis

Traditional:

# Manual EDA
import matplotlib.pyplot as plt
import seaborn as sns

# Create multiple plots manually
fig, axes = plt.subplots(2, 2, figsize=(15, 10))

# Sales over time
axes[0,0].plot(df['date'], df['sales'])
axes[0,0].set_title('Sales Trend')

# Sales by category
df.groupby('category')['sales'].sum().plot(kind='bar', ax=axes[0,1])
axes[0,1].set_title('Sales by Category')

# ... and 20 more visualizations

With AI:

Prompt: "Perform comprehensive EDA on this sales dataset, focus on trends, seasonality, and category performance"

AI Output:
- 15 relevant visualizations generated automatically
- Key patterns identified
- Anomalies highlighted
- Statistical summaries
- Correlation analysis
Time: 5 menit vs 2 jam traditional

3. Report Generation

Traditional:

  • Open PowerPoint
  • Create slides manually
  • Insert charts one by one
  • Write explanations
  • Format everything
  • Time: 4-5 jam

With AI:

Prompt: "Create executive summary presentation from this analysis, highlight top 3 insights, include visualizations"

AI generates:
- Structured presentation
- Key insights highlighted
- Charts embedded
- Draft narrative

Human refines:
- Add business context
- Adjust recommendations
- Customize for audience
- Total time: 1 jam

Real Productivity Gains: Case Study

Company: Mid-size E-commerce Startup

Before AI Implementation:

  • 1 analyst handle 2-3 projects/month
  • Average time per analysis: 2 weeks
  • Report quality: Good, but slow delivery

After AI Tools Integration:

  • Same analyst handle 6-8 projects/month
  • Average time per analysis: 3-4 days
  • Report quality: Better (AI catch errors humans miss)
  • 3x productivity increase

What Changed:

  • Data cleaning automated: 80% time saved
  • Initial analysis AI-generated: 70% time saved
  • More time for strategic thinking
  • Faster iteration based on stakeholder feedback

Quality Improvement, Not Just Speed

AI bukan cuma bikin cepat, tapi also lebih accurate:

Consistency:

  • Humans capek, make mistakes
  • AI konsisten analyze 10 rows atau 10 million rows
  • Standardized approach

Comprehensive Coverage:

  • Humans might miss patterns karena fatigue
  • AI check semua variables systematically
  • Better at spotting subtle correlations

Error Detection:

  • AI bisa catch logical inconsistencies
  • Validate calculations automatically
  • Highlight anomalies yang might be bugs

Example:

Manual analysis, analyst might miss:

  • Typo di SQL query (select sum instead of average)
  • Data from wrong time period
  • Sampling bias

AI tools automatically:

  • Validate query logic
  • Check date ranges consistency
  • Flag potential biases

Tools yang Bikin Perbedaan

For Data Cleaning & Prep:

  • ChatGPT/Claude: Generate cleaning scripts
  • DataRobot: Automated data prep
  • Trifacta: AI-powered data wrangling

For Analysis:

  • Pandas-AI: Natural language data analysis
  • Julius AI: Conversational analytics
  • ChatGPT Code Interpreter: Quick analysis

For Visualization:

  • Tableau Pulse: AI-generated insights
  • Power BI Copilot: Automated dashboards
  • Looker with AI: Smart visualizations

For Reporting:

  • Notion AI: Draft reports
  • Gamma: AI-powered presentations
  • Claude/ChatGPT: Narrative generation

Reality Check

AI Makes You Faster IF:

  • Lo tau apa yang lo cari (right questions)
  • Lo bisa validate AI outputs (catch errors)
  • Lo understand business context (apply insights correctly)

AI Won't Help IF:

  • Lo expect magic (garbage in = garbage out)
  • Lo ga check outputs (AI bisa salah)
  • Lo ga punya fundamental skills (blind reliance berbahaya)

The Pattern:

Without AI:
100% human effort → Slower, manual process → Good results

With AI (wrong approach):
100% AI, no human validation → Fast, but potentially wrong → Dangerous

With AI (right approach):
AI handles 60% grunt work + Human 40% judgment → Fast AND accurate → Best results

AI bikin lo lebih produktif, tapi lo tetep the pilot. AI tool yang powerful, bukan replacement. Next, kita bahas kenapa pilot (human) itu irreplaceable.

Bagian 4: AI Cuma Tools, Pilotnya Tetap Data Analyst

Okay, jadi AI bikin kerjaan lebih cepat. Cool. Tapi kenapa sih human tetep essential? Kenapa ga bisa full automation aja?

Let me explain dengan analogi yang relate.

Analogi yang Mudah Dipahami

AI itu kayak GPS buat driver:

  • GPS kasih route terbaik
  • GPS update real-time traffic
  • GPS warn kalau ada bahaya
  • Tapi yang nyetir tetep lo

GPS bisa error (nyuruh masuk gang buntu), GPS ga tau kalau ada hajatan yang tutup jalan, GPS ga bisa baca situasi. Driver yang decide.

AI kayak autocorrect buat penulis:

  • Autocorrect benerin typo
  • Suggest kata yang lebih baik
  • Check grammar
  • Tapi konten, storytelling, voice - that's still the writer

AI kayak kalkulator buat akuntan:

  • Kalkulator hitung cepat dan akurat
  • Automasi formula
  • Reduce human error
  • Tapi interpretasi, strategy, audit - that's the accountant

Pattern-nya sama untuk Data Analyst:

  • AI process data cepat
  • AI generate basic insights
  • AI buat visualizations
  • Tapi context, judgment, decisions - that's the analyst

Kenapa Human Tetap Essential

1. Context Understanding - AI Ga Paham "Why"

Real Case:

Data shows: Sales drop 25% minggu ini

AI Analysis:
- Sales: -25%
- Trend: Declining
- Recommendation: Increase marketing budget
- Confidence: 85%

Sounds logical, right?

Human Analyst Investigate:

  • Check calendar: Oh, ini minggu Lebaran
  • Normal behavior: Sales always drop saat long holiday
  • Context: Bukan masalah product/marketing
  • Decision: No action needed, wait for post-holiday rebound

AI miss context, human understand business cycles.

Another Example:

AI detects: Traffic spike 300% jam 2 pagi

AI flags: Potential DDOS attack
AI recommends: Block traffic source

Human checks:
- Traffic dari Indonesia
- Timing match dengan scheduled email campaign
- Newsletter sent jam 1 pagi (oops, wrong scheduling)
- Real users, bukan attack

Decision: Adjust email schedule, no blocking needed

AI sees patterns, human understands reasons.

2. Asking the Right Questions - Curiosity Beats Algorithms

AI cuma jawab apa yang lo tanya. Tapi good analyst tau pertanyaan apa yang seharusnya ditanya.

Example:

Stakeholder ask: "Why is our app rating dropping?"

Bad Analyst (rely fully on AI):

  • Input question ke AI
  • Get answer: "Ratings dropped from 4.5 to 4.2"
  • Present finding: "Rating turun 0.3 points"
  • Stakeholder: "Yeah I know, but WHY?"

Good Analyst (use AI as tool):

  • Use AI to pull rating data
  • Ask follow-up: "When did it start dropping?"
  • Segment: "Which platform? Android or iOS?"
  • Deep dive: "What features got negative reviews?"
  • Cross-check: "Any recent app updates?"
  • Find root cause: "Update 2.1.0 broke push notifications on Android"

The difference:

  • AI answers questions literally
  • Human asks the questions that matter
  • Human connects dots across different data sources

Business Context Example:

Question: "Should we discount Product A?"

AI calculation:
- Current profit margin: 30%
- 20% discount = margin drop to 10%
- Sales might increase 40%
- Net revenue: +5%
- Recommendation: Yes, do discount

Human analysis adds:
- Product A is premium brand positioning
- Discount might damage brand perception
- Competitors don't discount similar products
- Customer acquisition via discount = lower lifetime value
- Strategic decision: Maintain premium positioning
- Alternative: Bundle with complementary product

Decision: No discount, try bundling strategy

AI optimize for numbers, human optimize for long-term business value.

3. Stakeholder Management - EQ Beats IQ

Data analyst bukan cuma ngomong sama data. Lo ngomong sama people.

Scenario: Presenting to C-Level Executives

AI-generated report:

Sales Analysis Q4 2025

Key Metrics:
- Revenue: $2.5M (-5% QoQ)
- Conversion: 2.3% (-0.3pp)
- CAC: $45 (+12%)
- LTV: $180 (-8%)

Correlation matrix shows inverse relationship
between CAC and conversion (r=-0.72, p<0.05)

Regression analysis indicates...
[500 words of technical jargon]

Human analyst presents:

"We have a problem, but it's fixable.

Our revenue dipped 5% this quarter. Here's why:

[Visual: Simple bar chart showing trend]

The issue isn't demand - traffic is up 20%.
The issue is our checkout flow. We lost customers at payment step.

[Screenshot: Actual user frustration points]

Good news: Engineering can fix this in 2 weeks.
Expected impact: Get that 5% back + 3% more.

Investment needed: 1 sprint, ~$50k
Potential return: $400k/year

Question: Should we prioritize this for next sprint?"

The difference:

  • AI gives data
  • Human tells a story
  • Human understands what CEO cares about (ROI, timeline, risk)
  • Human presents in decision-making format

Managing Different Stakeholders:

Marketing Team:

  • Wants: "Which campaign works best?"
  • AI output: Statistical significance tables
  • Human translation: "Campaign B costs 30% less and gets 2x conversions"

Engineering Team:

  • Wants: Technical deep dive
  • AI output: High-level patterns
  • Human provides: Detailed hypothesis, data for debugging

Finance Team:

  • Wants: Bottom-line impact
  • AI output: Various metrics
  • Human focuses: Revenue, cost, ROI

Same data, different storytelling.

4. Ethical Considerations - Moral Compass Required

AI ga punya moral judgment. Human does.

Case 1: Bias Detection

AI Model Recommendation:
"Target ads for credit cards to high-income zip codes"

Analyst notices:
- These zip codes predominantly wealthy, white neighborhoods
- Excludes diverse communities
- Potentially discriminatory

Human decision:
- Adjust targeting to be income-based, not geography-based
- Ensure fair access to financial products

Case 2: Privacy Concerns

AI suggests:
"Track user behavior across devices using fingerprinting"

Data available:
- Browsing history
- App usage
- Location data
- Personal messages

Analyst considers:
- Is this necessary for business goal?
- Does user consent cover this?
- What's the privacy risk?

Decision:
- Use less invasive methods
- Be transparent with users
- Comply with regulations (GDPR, etc.)

Case 3: Algorithmic Fairness

AI-powered loan approval system:
- Approval rate: 85%
- Accuracy: 92%
- Looks good, right?

Analyst digs deeper:
- Approval rate for Group A: 95%
- Approval rate for Group B: 60%
- Model biased based on historical data

Action:
- Retrain model with fairness constraints
- Audit for demographic parity
- Implement bias monitoring

AI optimize for accuracy. Human ensure fairness and ethics.

5. Creative Problem Solving - Thinking Outside the Box

AI trained on existing patterns. Innovation requires thinking beyond patterns.

Real Case: Retention Problem

Standard approach (AI suggests):

  • Churn prediction model
  • Target at-risk users
  • Send discount coupon

Creative approach (human thinks):

  • Why are users churning?
  • Interviews: "Product is fine, I just forgot it exists"
  • Root cause: Engagement, not value
  • Solution: Gamification + social features
  • Result: 40% retention improvement (better than discounts)

AI suggests what worked before. Human invents what works next.

Another Example: Data Storytelling

Boring (AI-generated):
"Sales increased 25% in Q4 due to promotional campaigns and seasonal factors"

Engaging (Human-crafted):
"Remember when we launched that 12.12 campaign? Everyone was skeptical.
'Too much discount,' they said. 'We'll lose margin.'

Here's what actually happened:
[Visual journey showing campaign progression]

Not only did we hit our target, we exceeded it by 25%.
But here's the interesting part - 60% of those customers came back in January.
Without any discount.

That's the power of right timing + right offer."

Same data, completely different impact.

Limitations of AI in Data Analysis

Let's be real about AI's weaknesses:

1. Hallucinations

  • AI bisa generate insights yang "sounds smart" tapi salah
  • Confidence doesn't mean correctness
  • Example: ChatGPT might say "correlation is 0.85" when actual is 0.25

2. No Common Sense

AI sees: Ice cream sales correlate with drowning deaths
AI concludes: Ice cream causes drowning
Human knows: Both increase in summer, correlation ≠ causation

3. Can't Handle Novel Situations

  • AI trained on historical data
  • Pandemic? Economic crisis? New competitor?
  • AI struggling dengan unprecedented events
  • Human adapts

4. No Emotional Intelligence

  • Can't read room
  • Can't gauge stakeholder reactions
  • Can't adjust communication style
  • Can't build relationships

5. Requires Perfect Instructions

Prompt: "Analyze sales data"

AI needs:
- What time period?
- Which products?
- What metrics matter?
- How to handle outliers?
- What format output?

AI can't fill in the blanks. Human can.

6. Context Window Limitations

  • AI process limited amount of information
  • Complex projects with 50+ data sources?
  • Long-term business context from 5 years ago?
  • Human brain connects dots across time and sources

The Right Relationship: Human + AI

Think of it like this:

AI = Super smart intern
- Works 24/7
- Never complains
- Fast execution
- Needs clear direction
- Requires validation

Human = Senior analyst
- Provides direction
- Asks right questions
- Validates outputs
- Adds context
- Makes final decisions

Workflow yang Bener:

1. Human: Define problem
   "Why is retention dropping?"

2. AI: Generate initial analysis
   Pull data, calculate metrics, identify patterns

3. Human: Validate & dig deeper
   Check AI work, ask follow-ups, cross-reference

4. AI: Execute deep dives
   Segment analysis, statistical tests, visualizations

5. Human: Synthesize & strategize
   Combine findings, add business context, recommend actions

6. AI: Document & automate
   Create report template, set up monitoring

7. Human: Present & decide
   Stakeholder communication, final recommendations

Real Success Story:

Company: E-learning Platform

Challenge: User engagement dropping

AI's role:

  • Analyze 50 million user interactions
  • Identify drop-off points
  • Segment user behaviors
  • Generate 100+ hypotheses
  • Time: 2 hours

Human's role:

  • Filter to 5 actionable hypotheses
  • Understand why patterns exist
  • Design experiments
  • Communicate to product team
  • Monitor results and iterate
  • Time: 1 week, but strategic work

Outcome:

  • 35% engagement increase
  • Feature priorities refined
  • Product roadmap adjusted

Could AI do it alone? No. Too much judgment needed. Could human do it alone? Yes, but would take 3 months instead of 1 week. Together? Best of both worlds.

Your Role in 2026: The Pilot

Lo bukan cuma "data analyst" lagi. Lo adalah:

AI-Augmented Analyst

  • Fluent in AI tools
  • Can direct AI effectively
  • Validate AI outputs
  • Fill in AI gaps with human insight

Think of yourself as:

  • Conductor of AI orchestra (AI plays instruments, you direct the symphony)
  • Editor of AI outputs (AI writes draft, you perfect it)
  • Strategist with AI assistant (AI provides data, you provide strategy)

Your competitive advantage:

  • Not "I can do what AI can't"
  • But "I can leverage AI + add value AI can't"

The future belongs to analysts who:

  1. Embrace AI tools (not resist)
  2. Develop uniquely human skills (judgment, communication, creativity)
  3. Focus on strategic work (let AI handle grunt work)
  4. Stay curious and keep learning

AI won't replace you. But data analysts who use AI will replace those who don't.

Next, let's talk about konkret skill evolution dan action plan.

Bagian 5: Skill Evolution & Action Plan untuk Data Analyst 2026

Okay, sekarang lo udah paham: AI won't replace you, but you need to evolve. Question-nya: evolve gimana? Start dari mana?

Let me break it down konkret.

Skill Set yang Perlu di-Upgrade

Technical Skills (The Foundation)

1. AI Literacy (MUST HAVE - Non-negotiable)

Bukan berarti lo harus bisa bikin AI model dari scratch. Maksudnya:

Understand How AI Works:

  • Basic ML concepts (supervised vs unsupervised learning)
  • Know AI limitations (hallucinations, bias, context window)
  • Understand when AI appropriate vs overkill

Prompt Engineering:

`Bad prompt: "Analyze this data"

Good prompt: "Analyze sales data from Q4 2025. Focus on:

  1. Month-over-month trends
  2. Top 5 performing products by revenue
  3. Regional performance comparison
  4. Identify any anomalies or outliers
  5. Suggest potential drivers for changes

Present findings in executive summary format with visualizations."`

Validate AI Outputs:

  • Cross-check AI calculations
  • Question suspicious insights
  • Test with known datasets
  • Don't blindly trust

Tools to Master:

  • ChatGPT/Claude for analysis assistance
  • GitHub Copilot for code generation
  • AI-powered BI tools (Tableau Pulse, Power BI Copilot)

2. Programming (Enhanced & Essential)

Python - Your Best Friend:

Core libraries:

python

`# Data manipulation import pandas as pd import numpy as np

Visualization

import matplotlib.pyplot as plt import seaborn as sns import plotly.express as px

Statistical analysis

from scipy import stats import statsmodels.api as sm

Machine learning basics

from sklearn.model_selection import train_test_split from sklearn.preprocessing import StandardScaler from sklearn.linear_model import LinearRegression

AI integration

import openai # For API calls from pandasai import SmartDataframe`

Real-world skills:

  • Data wrangling (merge, join, pivot, reshape)
  • Time series analysis
  • A/B test analysis
  • Automated reporting scripts

SQL - Still King:

sql

  • - Advanced SQL skills neededWITH monthly_metrics AS ( SELECT DATE_TRUNC('month', order_date) as month, user_id, COUNT(order_id) as order_count, SUM(revenue) as total_revenue FROM orders WHERE status = 'completed' GROUP BY 1, 2),user_segments AS ( SELECT user_id, CASE WHEN order_count >= 5 THEN 'High frequency' WHEN order_count >= 2 THEN 'Medium frequency' ELSE 'Low frequency' END as segment FROM monthly_metrics )SELECT s.segment, COUNT(DISTINCT s.user_id) as user_count, AVG(m.total_revenue) as avg_revenue, PERCENTILE_CONT(0.5) WITHIN GROUP (ORDER BY m.total_revenue) as median_revenue FROM user_segments s JOIN monthly_metrics m ON s.user_id = m.user_id GROUP BY 1ORDER BY avg_revenue DESC;

Level up:

  • Window functions
  • CTEs (Common Table Expressions)
  • Query optimization
  • Working with large datasets (millions of rows)

R (Optional but Valuable):

  • Statistical modeling
  • Specialized packages for specific analysis
  • Academic/research environments prefer R

Version Control (Git):

bash

`# Essential commands git init git add . git commit -m "Add customer segmentation analysis" git push origin main

Collaboration

git pull git branch feature/new-analysis git merge`

Why? Collaboration, reproducibility, portfolio.

3. Machine Learning Basics (No Need to be Expert)

Lo ga perlu jadi ML engineer. Tapi perlu understand concepts:

Know These Algorithms:

  • Linear/Logistic Regression
  • Decision Trees/Random Forest
  • K-means Clustering
  • Time Series Forecasting (ARIMA, Prophet)

Understand:

  • Train/test split
  • Overfitting vs underfitting
  • Model evaluation metrics (accuracy, precision, recall, RMSE)
  • Cross-validation

When to Use ML:

`Good use cases: ✅ Predict customer churn ✅ Forecast demand ✅ Segment customers ✅ Anomaly detection

Bad use cases (overkill): ❌ Calculate average sales (use SQL) ❌ Simple trend analysis (use Excel) ❌ One-time exploratory analysis`

Simple Example:

python

`from sklearn.ensemble import RandomForestClassifier from sklearn.metrics import classification_report

Predict churn

X = df[['tenure', 'monthly_charges', 'total_charges']] y = df['churn']

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)

model = RandomForestClassifier() model.fit(X_train, y_train)

predictions = model.predict(X_test) print(classification_report(y_test, predictions))`

Lo ga perlu tau semua detail math-nya. Tapi perlu tau:

  • Kapan use case-nya
  • How to implement
  • How to interpret results

4. Cloud Platforms (Industry Standard)

Google BigQuery:

sql

  • - Analyze terabytes of data in secondsSELECT user_pseudo_id, COUNT() as events, MAX(event_timestamp) as last_seen FROM project.analytics_123456789.events_*WHERE _TABLE_SUFFIX BETWEEN '20250101' AND '20250131'GROUP BY 1HAVING events > 10

AWS (Amazon Web Services):

  • S3: Data storage
  • Athena: Query data in S3
  • Redshift: Data warehousing

Azure:

  • Azure Synapse Analytics
  • Power BI integration

Why Cloud?

  • Scale (handle huge datasets)
  • Collaboration (shared data/queries)
  • Cost-effective
  • Industry standard

Soft Skills (Increasingly Critical)

1. Storytelling - Make Data Speak

Bad presentation:

"Revenue increased 15% QoQ Customer count up 20% Churn rate down 3% AOV stable at $45"

Good storytelling:

`"Last quarter, we made a bet on retention over acquisition.

[Visual: Investment shift from ads to product improvement]

The results? Not only did we keep more customers (churn down 3%), but they brought their friends. 20% more customers, 15% more revenue.

Best part? Our cost per acquisition actually decreased 25%.

Here's what we learned and what's next..."`

Elements of Good Storytelling:

  • Start with the "so what"
  • Use visuals that support narrative
  • Connect to business goals
  • Provide actionable next steps

2. Business Acumen - Think Like a Business Owner

Understand Your Company:

  • What are the key revenue drivers?
  • What metrics does leadership obsess over?
  • How does your team make money?
  • Who are the competitors?

Link Analysis to Business:

Weak: "Engagement rate is 25%" Strong: "Our 25% engagement translates to $500k monthly revenue. A 5% improvement would mean $100k additional revenue. Here's how we get there..."

Think ROI:

  • Every analysis should answer: "So what should we do?"
  • Every recommendation should include: Expected impact + Required resources

3. Communication - Tailor Your Message

For Technical Audience (Engineers, Data Scientists):

  • Detail-oriented
  • Show methodology
  • Discuss assumptions and limitations
  • Technical terminology OK

For Business Audience (Product, Marketing):

  • Focus on insights and actions
  • Less technical jargon
  • More "what it means for us"
  • Clear recommendations

For Executive Audience (C-level):

  • Start with conclusion
  • High-level only
  • Focus on business impact
  • Be ready for tough questions

Master Different Formats:

  • Quick Slack update: "Conversion up 5%, driven by mobile. Details in dashboard."
  • Email summary: 3-5 bullet points, link to full report
  • Presentation: 10 slides max, tell a story
  • Dashboard: Self-service, clear labels, intuitive

4. Critical Thinking - Question Everything

Develop Healthy Skepticism:

`Data shows correlation between ice cream sales and drowning.

Bad analyst: "Ice cream causes drowning!"

Good analyst:

  • What's the time period? (Both peak in summer)
  • Any confounding variables? (Temperature, season)
  • Does causation make sense? (No)
  • Real insight: Seasonal pattern, not causation`

Challenge AI Outputs:

  • Does this make business sense?
  • Are there potential biases?
  • What's missing from this analysis?
  • What alternative explanations exist?

Question Stakeholder Requests:

`Stakeholder: "Show me conversion rate by day of week"

Don't just execute. Ask:

  • Why do you need this? (Understand goal)
  • What decision will this inform? (Ensure relevance)
  • What's the time period? (Set scope)
  • Any specific hypotheses? (Guide analysis)`

5. Adaptability - Embrace Change

Stay Current:

  • New tools emerge monthly
  • Best practices evolve
  • Technologies change

Be Comfortable with Uncertainty:

  • Not every analysis has clear answer
  • Sometimes data is messy
  • Some questions are unanswerable

Learn Continuously:

  • Follow industry blogs
  • Join communities
  • Take courses
  • Experiment with new tools

Action Plan Konkret

Phase 1: Foundation (Bulan 1-2)

Week 1-2: SQL Mastery

  • [ ] Complete SQL fundamentals
  • [ ] Practice JOIN operations
  • [ ] Master aggregations and GROUP BY
  • [ ] Learn window functions
  • Resource: BuildWithAngga - "Belajar SQL dari Nol" (GRATIS)

Week 3-4: Excel/Google Sheets Advanced

  • [ ] Pivot tables mastery
  • [ ] Advanced formulas (VLOOKUP, INDEX/MATCH)
  • [ ] Data cleaning techniques
  • [ ] Basic visualizations
  • Resource: BuildWithAngga - "Excel untuk Data Analysis"

Week 5-6: Statistics Fundamentals

  • [ ] Descriptive statistics
  • [ ] Probability basics
  • [ ] Hypothesis testing
  • [ ] Correlation vs causation

Week 7-8: First Project

  • [ ] Choose dataset (Kaggle, company data, etc.)
  • [ ] Perform complete analysis
  • [ ] Create presentation
  • [ ] Publish on GitHub/LinkedIn

Phase 2: Technical Deep Dive (Bulan 3-4)

Week 1-4: Python for Data Analysis

  • [ ] Python basics
  • [ ] Pandas for data manipulation
  • [ ] NumPy for numerical computing
  • [ ] Matplotlib & Seaborn for viz
  • Resource: BuildWithAngga - "Python untuk Data Analyst" (GRATIS)
    • 30+ video modules
    • Real-world projects
    • Portfolio-ready outputs
    • Lifetime access

Week 5-8: Visualization Tools

  • [ ] Tableau fundamentals
  • [ ] Dashboard design principles
  • [ ] Interactive visualizations
  • [ ] OR Power BI (choose one)
  • Resource: BuildWithAngga - "Tableau untuk Pemula"
    • Step-by-step tutorials
    • Dashboard templates
    • Best practices Indonesia

Mini Projects:

  • E-commerce sales analysis
  • Customer segmentation
  • Cohort analysis
  • Funnel analysis

Phase 3: AI Integration (Bulan 5-6)

Week 1-2: AI Literacy

  • [ ] Understand how ChatGPT/Claude works
  • [ ] Learn prompt engineering
  • [ ] Practice with data analysis prompts
  • [ ] Validate AI outputs

Week 3-4: Machine Learning Basics

  • [ ] Scikit-learn fundamentals
  • [ ] Classification basics
  • [ ] Clustering techniques
  • [ ] Model evaluation
  • Resource: BuildWithAngga - "Intro to Machine Learning"

Week 5-6: Integrate AI into Workflow

  • [ ] Use ChatGPT for code generation
  • [ ] AI-assisted data cleaning
  • [ ] Automated report generation
  • [ ] Compare AI vs traditional methods

Capstone Project:

  • Complex analysis using AI + traditional methods
  • Document your process
  • Show AI collaboration
  • Present insights

Phase 4: Continuous Learning (Ongoing)

Daily (15-30 min):

  • Read one data analytics article
  • Practice SQL on LeetCode/HackerRank
  • Explore dataset on Kaggle

Weekly (2-3 hours):

  • Complete mini analysis project
  • Write about learnings (LinkedIn/blog)
  • Experiment with new tool

Monthly:

  • Attend webinar or workshop
  • Network with other analysts
  • Review and update portfolio
  • Learn one new technique

Quarterly:

  • Take advanced course
  • Contribute to open source
  • Speak at meetup/write tutorial
  • Reassess career goals

Rekomendasi Kelas BuildWithAngga

🎯 Kelas GRATIS untuk Mulai

1. "Belajar SQL dari Dasar: Zero to Hero"

  • Durasi: 8 jam video
  • Level: Pemula sampai intermediate
  • Benefit:
    • Kuasai SQL query dasar sampai advanced
    • Window functions dan CTEs
    • Real-world case studies
    • Practice dengan database sebenarnya
    • Sertifikat completion
  • Project: Analisis data e-commerce lengkap
  • Link: [buildwithangga.com/kelas/sql-fundamental]

2. "Python untuk Data Analyst"

  • Durasi: 12 jam hands-on
  • Level: Pemula friendly
  • Benefit:
    • Python basics sampai data analysis
    • Pandas, NumPy, Matplotlib mastery
    • Data cleaning, EDA, visualization
    • 5 mini projects portfolio-ready
    • Lifetime access
    • Code templates
  • Bonus: Jupyter notebook templates, dataset collection
  • Link: [buildwithangga.com/kelas/python-data-analyst]

3. "Excel untuk Data Analysis"

  • Durasi: 6 jam praktik
  • Benefit:
    • Advanced Excel techniques
    • Pivot tables mastery
    • Data visualization
    • Dashboard creation
    • Real business cases
  • Perfect untuk: Beginners yang mau quick wins
  • Link: [buildwithangga.com/kelas/excel-analysis]

🚀 Kelas Premium (Investment Worthy)

4. "Complete Data Analyst Bootcamp"

  • Harga: Rp 499k (one-time, lifetime access)
  • Durasi: 40+ jam comprehensive
  • Curriculum:
    • SQL advanced
    • Python data analysis
    • Tableau/Power BI
    • Statistics for analysts
    • Machine learning basics
    • Portfolio building
  • Includes:
    • Mentorship session
    • Career guidance
    • Job hunting prep
    • Community access
    • Updated content regular

5. "Business Intelligence dengan Tableau"

  • Harga: Rp 299k
  • Benefit:
    • Professional dashboard creation
    • Storytelling with data
    • Interactive visualizations
    • Real client projects
    • Portfolio showcase

💎 Kenapa BuildWithAngga?

1. Bahasa Indonesia:

  • Ga perlu struggle dengan English
  • Context Indonesia (currency, examples)
  • Relatable case studies

2. Project-Based Learning:

  • Bukan cuma teori
  • Langsung practice
  • Build portfolio sambil belajar

3. Lifetime Access:

  • Bayar sekali, akses selamanya
  • Course updates included
  • Re-watch kapanpun

4. Community Support:

  • Forum diskusi
  • Telegram group
  • Mentor support
  • Network dengan sesama learners

5. Career Focus:

  • CV review
  • Portfolio tips
  • Interview prep
  • Job hunting guidance

6. Affordable:

  • Free courses berkualitas
  • Premium courses worth the investment
  • Often ada diskon (monitor website)

Skill Evolution Roadmap

`Month 1-2: FOUNDATION ├─ SQL Basics ✓ ├─ Excel Advanced ✓ ├─ Statistics Fundamentals ✓ └─ First Portfolio Project ✓

Month 3-4: TECHNICAL DEPTH ├─ Python Data Analysis ✓ ├─ Data Visualization (Tableau) ✓ ├─ Multiple Mini Projects ✓ └─ GitHub Portfolio ✓

Month 5-6: AI INTEGRATION ├─ AI Literacy & Prompting ✓ ├─ ML Basics ✓ ├─ AI-Assisted Workflow ✓ └─ Capstone Project ✓

Month 7+: SPECIALIZATION ├─ Choose domain (Marketing/Product/Finance Analytics) ├─ Advanced techniques ├─ Thought leadership (blogs, talks) └─ Continuous learning`

Career Paths Available

Traditional Path (Enhanced by AI):

Junior Data Analyst (0-2 years) ↓ Mid-Level Data Analyst (2-4 years) ↓ Senior Data Analyst (4-7 years) ↓ Lead/Principal Analyst (7+ years)

AI-Native Roles:

`AI-Assisted Data Analyst ├─ ML Analyst ├─ AI Product Analyst └─ Analytics Engineer

Data Science Track ├─ Junior Data Scientist └─ ML Engineer`

Specialized Tracks:

`Domain Expertise ├─ Marketing Analytics Specialist ├─ Product Analytics Specialist ├─ Financial Analyst └─ Operations Analyst

Technical Leadership ├─ Analytics Manager ├─ Head of Analytics └─ Chief Data Officer`

Emerging Roles (2026+):

  • AI Ethics Analyst: Ensure fair AI use in analytics
  • Prompt Engineer for Analytics: Optimize AI for data tasks
  • Human-AI Collaboration Specialist: Bridge between AI and business
  • Analytics Product Manager: Build AI-powered analytics tools

Salary Expectations (Indonesia, 2026)

Jakarta/Major Cities:

  • Junior (0-2 yr): Rp 7-12 juta/bulan
  • Mid-level (2-4 yr): Rp 12-20 juta/bulan
  • Senior (4-7 yr): Rp 20-35 juta/bulan
  • Lead/Principal (7+ yr): Rp 35-60 juta/bulan

With AI Skills Premium:

  • Add 20-30% to base salary
  • More negotiation leverage
  • Access to better companies

Remote International:

  • USD $2,000-5,000/month (junior-mid)
  • USD $5,000-10,000/month (senior)

Final Tips

1. Start Now, Don't Wait for Perfect

  • Ga perlu tunggu "ready"
  • Learning by doing
  • Mistakes = lessons

2. Build in Public

  • Share progress di LinkedIn
  • Write about learnings
  • Help others = solidify knowledge

3. Focus on Projects, Not Certificates

  • Portfolio > Certificates
  • Real work > Online badges
  • Show, don't tell

4. Network Actively

  • Join communities
  • Attend meetups
  • Connect with practitioners
  • Learn from others' journey

5. Stay Curious

  • Ask "why" banyak-banyak
  • Experiment with new tools
  • Challenge assumptions
  • Never stop learning

Remember: AI won't replace data analysts. Data analysts who use AI will replace those who don't.

The future is collaborative. Be the analyst who embraces AI, develops human skills, and creates value that no AI alone can deliver.

Penutup: The Future is Yours

Jadi, apakah AI akan gantikan data analyst di 2026?

No. Absolutely not.

But AI WILL change:

  • How you work (faster, more efficient)
  • What skills matter (communication, judgment > technical execution)
  • What you focus on (strategy > grunt work)

The opportunity:

Ini literally best time untuk jadi data analyst:

  • High demand
  • Good salary
  • Powerful tools available
  • Learning resources abundant
  • Remote work options

Your action items:

✅ This Week:

  • Daftar kelas gratis BuildWithAngga
  • Download first dataset
  • Start learning SQL/Python

✅ This Month:

  • Complete first mini project
  • Create LinkedIn profile
  • Join data analytics community

✅ This Quarter:

  • Finish foundation courses
  • Build 3-5 portfolio projects
  • Start job applications

✅ This Year:

  • Land data analyst role
  • Master AI tools
  • Continuously upgrade skills

Remember:

"The best time to start was yesterday. The second best time is now."

AI is here to stay. Data analysts are here to stay. The question bukan "will I survive?" tapi "how will I thrive?"

Start your journey today. Literally hari ini.

[CALL TO ACTION]

🎯 Siap Mulai? Join BuildWithAngga Gratis:

  • 👉 [SQL Fundamental Course - GRATIS]
  • 👉 [Python untuk Data Analyst - GRATIS]
  • 👉 [Join Community Discord]

💬 Share Your Journey: Comment di bawah:

  • Udah pernah pake AI tools buat analysis?
  • Skill apa yang mau lo upgrade duluan?
  • Challenge terbesar lo soal AI?

Let's learn together. The future of data analytics is bright, dan lo bisa jadi bagian dari itu.