Data Architecture

Question 1

Define Data Architecture

Answer 1

A framework 🏗️ that

defines how data is

collected,📥
stored, 🗄️
managed, 📋
and accessed 🔑

across an organisation

Designed with business need in mind

Question 2

Name 3 tools that you work with as part of you data systems architecture

Answer 2

🟢 Collibra - metadata on JLP’s data assets and RoP

🔤 Alfabet - records the applications used in JLP and how they are connected - the IT landscape

❄️ Snowflake - cloud based storage and analytics service that handles both structured and unstructured data

🟦 Tableau - JLP tool of choice for data visualisation used for reporting

Question 3

Why did you use regression analysis to investigate your data?

Answer 3

• Pearson’s Correlation identified a weak + relationship between Lead Time & Stock Holding
• Regression would show me the linear relationship to understand how stock holding varies with lead time
• Forecast stock holding based on lead time

Question 4

Why did you use time series forecasting to investigate you data?

Answer 4

• Regression Model:
• not accurate;
• other variables;
• weak relationship
• Time Series Forecast to identify patterns in Stock Holding and forecast using Historical data

Question 5

How did you access data from your organisation’s data system and how did this impact your tool choice

Answer 5

• Collibra/Alfabet Direct Extract
• Google Sheets
• Tableau: Integrated with GS

Question 6

Why did you use Tableau?

Answer 6

1. Integrates w/ **GS **🗒️
2. Interactive🎮
3. Secure sharing - Tableau Online, accessible from any device 📱
4. Stakeholders familiar with it - JLP** tool of choice** ✅
5. User-friendly join feature 🤝
6. Future automation: Snowflake or Collibra Application Programming Interace 🤖

Question 7

Why did you use GS?

Answer 7

1. User friendly interface - formulas ➗
2. JLP tool of choice ✅
3. Collaboration 🤝
4. Tableau Integration/File Import 🔗

Question 8

Why did you use Python? (Hypothesis Testing)

Answer 8

For my hypothesis test (stock holding and lead time relationship):
1. built in libraries
2. visualisation like scatter plot & hypothesis test in one tool
3. handles big data sets
4. reuse the same code for different sets of data

Question 9

Why did you use Python? (Regression)

Answer 9

1. Greater flexibility
2. Greater control
3. Tableau linear regression operates behind scenes
4. E.g. Tableau wouldn’t have split the data into test/train