Bangladesh District Population & Growth, 1991–2022
Thirty-one years of census data across all 64 districts, built in Tableau and published publicly — an interactive view of where Bangladesh actually grew.
This project can be shown in full. It uses a public dataset with no sensitive content, so every figure, method and result on this page is the real thing — the visuals are screenshots of the working output, not rebuilds.
Context
This was my first build in Tableau. I had been working almost entirely in Power BI, so a good part of the time went into learning how Tableau expresses the same ideas differently — particularly its handling of geography, which is considerably better out of the box.
The data covers all 64 districts of Bangladesh across 8 divisions, with land area and the population counted in four census years: 1991, 2001, 2011 and 2022. The question worth asking of it is not “how many people” but “where did the growth actually happen, and does district size have anything to do with it?”
Approach
The brief asked for four analytical questions. I wrote six, because four left gaps that a reader would notice:
- National trajectory across the four census years, to set the scale before any detail.
- Growth by division, 1991 to 2022 — which parts of the country are actually expanding.
- Top ten districts by 2022 population, for where people are concentrated today.
- City-size breakdown, grouping districts into mega / large / medium / small to show the balance between a few giants and many smaller places.
- Area against population as a scatter, to test the intuitive assumption that bigger districts hold more people.
- A choropleth map, so the numbers connect to real places rather than staying abstract.
The dashboard
Built on Kaggle’s Bangladesh District Population Census data (1991, 2001, 2011, 2022). Real screenshot — the live interactive version is linked above.
Outcome
170M across 64 districts
Growth is sharply uneven. Dhaka division grew 83% between 1991 and 2022, more than four times Barisal’s 20%, and Dhaka district alone now holds 15.21 million people — more than the next two districts combined. Nationally the population moved from 111M to 170M across the four censuses.
The scatter answered its question with a clear no: district area does not predict population. Rangamati is one of the largest districts by land area and one of the smallest by population, while Dhaka is the inverse. Density, not size, is what planning has to work with.
What I’d do differently
Coming from Power BI, my first instinct was to over-filter — I initially built separate views for each division before realising Tableau’s dashboard actions could do the same job from one canvas. The published version is considerably simpler than my first attempt, and better for it.
Analytically, four census points across 31 years is enough to show direction but not enough to model it. Adding annual population estimates between censuses would make growth-rate projection possible, which is what a planning audience would actually want next.