Statistics - lesson 8 | HackIndia Skills

Key point

Your proof bundle

A data-backed project is not just a chart in your demo. It is a small proof bundle: dataset or sample, chart, insight, limitation, and a link to the code that made it.

Make it easy to verify in under two minutes. Judges and recruiters should not need to DM you for the CSV.

Your to-do

Before you submit

Do each one yourself, then tap it to tick it off. The ticks are only a checklist for you: they are not marked or scored.

Save your dataset or sample in the repo, or link to it if it is too large or private.
Create one chart that directly supports the feature or analysis you are claiming.
Write one plain-language insight that a non-data person can understand.
Write one limitation that reduces overclaiming, such as sample size, bias, missing fields, or short test duration.
Add all four items to your README and hackathon submission description.

0 of 5 done

What counts

Your dataset can be small. It can be 50 survey responses, scraped public data, app usage logs from your demo, or a manually labelled sample.

Do not hide weak data. If your sample is only from your college WhatsApp groups, say that. Honest limits show judgement.

Example

README evidence block

Use a short section like this in your README. Replace the details with your own project data.

Evidence

Dataset or sample:
- File: data/feedback_sample.csv
- Rows: 83 student feedback responses collected during testing
- Columns: timestamp, college_year, task_completed, rating, comment

Chart:
- File: reports/task_completion_by_year.png
- What it shows: completion rate by college year

Plain-language insight:
- In our small test, second and third year students completed the onboarding task more often than first year students.

Limitation:
- The sample came from our own college network, so it may not represent students from other colleges or cities.

How to reproduce:
- Run: python scripts/make_chart.py

Tip

Use AI carefully

You can ask ChatGPT, Gemini, or Claude to draft your evidence section from your CSV summary. Give it row count, column names, chart title, and your rough finding.

Check the output for overclaiming. Replace words like “proves”, “always”, and “users prefer” with safer words like “suggests”, “in this sample”, and “we observed”.

Common mistake

Do not submit vibes

A sentence like “users loved it” is not a data-backed insight unless you show what data supports it.

Also avoid a lonely chart with no explanation. A chart needs a claim and a limitation, or people will fill in their own story.

Your to-do

Submission checklist

Do each one yourself, then tap it to tick it off. The ticks are only a checklist for you: they are not marked or scored.

Link your public repo in the HackIndia submission.
Include the dataset file, sample file, or data source note.
Include the chart image or notebook output in the repo.
Put the insight and limitation in the README, not only in your pitch deck.
Mention the same insight briefly in your demo video or submission text.

0 of 5 done

Your to-do

Do this now

Do each one yourself, then tap it to tick it off. The ticks are only a checklist for you: they are not marked or scored.

Open your project README and add a section called Evidence.
Add links to your dataset or sample and your chart file.
Write one insight using the phrase “In this sample”.
Write one limitation that a judge would care about.
Commit and push the README update to your public repo.

0 of 5 done

Submit The Data-Backed Project

Your proof bundle

Before you submit

What counts

README evidence block

Use AI carefully

Do not submit vibes

Submission checklist

Do this now

Quick check