Data Analytics Projects for Your Resume

Hiring managers for analyst roles look for evidence of three things: that you can get and clean data, that you can find something that matters in it, and that you can explain it to people who will act on it. A good portfolio project shows all three on a real question. The projects below use real datasets whose licences allow reuse, and each is framed as a question a business would actually ask. Pick two or three that suit the roles you want; together they should cover SQL, a BI tool and a written recommendation.

1. Sales and returns review for an online retailer

Data: a year of transactions from a UK online gift retailer, over 540,000 lines (UCI: Online Retail). Question: How is the business performing month by month, what drives revenue, and how much do returns cost? Skills shown: cleaning cancellations, fees and missing IDs; defining KPIs precisely; monthly trends; breakdowns by country and product. Deliverable: a one-page dashboard plus a short note on the three most important findings. Stretch: estimate how much revenue depends on the top 10 customers and what that concentration means.

Building a sales dashboard works through this project step by step.

2. Customer retention and value

Data: the same retailer's transactions, analysed at the customer level. Question: Do new customers come back, and which customers matter most? Skills shown: cohort retention tables, RFM segmentation, reading results critically, such as spotting a cohort distorted by the start of the data. Deliverable: a retention table, a segment summary with each segment's share of revenue, and one recommended action per segment. Stretch: track how customers moved between segments from one quarter to the next.

Cohort analysis explained and customer segmentation for analysts show the methods on this data.

3. Bike-share operations analysis

Data: two years of hourly bike rentals in Washington, D.C., with weather and calendar information (UCI: bike sharing). Question: When and in what conditions is demand highest, and what does that mean for staffing and bike redistribution? Skills shown: time-based analysis, comparing working days and weekends, weather effects, and avoiding confounded conclusions, since time of day drives both weather readings and rentals. Deliverable: an operations brief with demand patterns and two concrete recommendations. Stretch: a simple forecast of next week's daily demand, compared with a naive baseline.

4. A SQL business review of a music store

Data: the Chinook sample database, a digital media store with customers, employees, invoices, tracks and genres (Chinook database). Question: Which genres, countries and sales representatives drive revenue, and where is the store under-performing? Skills shown: multi-table joins, aggregation at the right grain, window functions for rankings and running totals, and reconciling totals after joins. Deliverable: a repository of well-commented SQL queries with their results and a one-page summary. Stretch: find the catalogue share that has never sold and suggest what to do about it.

5. Country development indicators report

Data: World Bank indicators such as GDP per capita, internet use and education, most released under a Creative Commons Attribution licence (World Bank: data licences). Question: How have internet access and related indicators changed across countries in a region over the last decade? Skills shown: reshaping wide indicator tables, handling missing years, comparing groups fairly, and writing about correlations without claiming causes. Deliverable: an illustrated report for a non-specialist reader. Stretch: an interactive dashboard with a country filter.

6. Developer survey analysis

Data: the Stack Overflow Developer Survey, whose data is released under the Open Database License (Stack Overflow Developer Survey). Question: Which tools and languages do developers in different roles and experience levels use, and how is that changing? Skills shown: cleaning multi-select survey answers, grouping and comparing respondents, and stating the limits of a self-selected sample. Deliverable: five charts, each with a one-sentence takeaway, and a note on who the survey does and does not represent. Stretch: compare two survey years to identify rising and declining tools.

7. An Indian public dataset of your choice

Data: a dataset from the Government of India's open data platform relevant to your city, state or industry (OGD Platform India). Check each dataset's licence and documentation before use. Question: one that a local business, NGO or government office would care about, such as how a service's usage differs between districts. Skills shown: finding and documenting public data, cleaning inconsistent administrative data, and writing for a real audience. Deliverable: a short report or dashboard addressed to that audience. Stretch: combine two datasets on a shared key, such as district names, and document how you matched them.

Matching projects to the roles you want

What makes a project stand out

• A real question and a clear answer. "Sales rose 101% from August to November, almost entirely from more orders" beats "I explored the data".
• Documented cleaning with row counts for each exclusion.
• Honest limits: what the data cannot tell you.
• A recommendation someone could act on.
• Clean presentation: a readable repository, a short write-up, and a dashboard screenshot or link.

How to build a data analyst portfolio covers how to present these projects to recruiters.

Projects on sales, customer segmentation and KPIs, built with Excel, SQL, Python and Power BI, are part of the Data Analytics course.

Sources

• Online Retail (UCI Machine Learning Repository) — UCI Machine Learning Repository
• Bike Sharing (UCI Machine Learning Repository) — UCI Machine Learning Repository
• Chinook Database (GitHub) — Luis Rocha
• World Bank data licences — World Bank
• Stack Overflow Developer Survey — Stack Overflow
• Open Government Data (OGD) Platform India — Government of India