Bangladesh Food Price Intelligence Dashboard
Twenty-six years of World Food Programme price data — 14,178 observations across 74 markets — turned into an interactive Power BI tool for tracking food inflation, regional disparity and supply-chain margin.
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
The World Food Programme has published food price observations for Bangladesh since 1998. The archive is large and awkward: 14,178 observations across 74 markets, 64 districts and 8 divisions, spanning July 1998 to August 2024, with prices recorded in different units and at both wholesale and retail level.
In that shape it answers nothing. The work was to make twenty-six years of it navigable enough that someone could ask a question — which region pays most, is the gap widening, does the exchange rate explain it — and get an answer in a few clicks.
Approach
- Normalise before anything else. Prices arrive in mixed units, so a
price_per_unitmeasure converts everything to a per-kilogram basis. Without that step every cross-commodity comparison on the page would be meaningless. - Five questions, five visuals. How prices moved over 26 years; which regions pay most; how wide the wholesale-to-retail gap is; whether currency depreciation explains inflation; and which season is most expensive.
- A middleman-margin measure. Retail minus wholesale, expressed as a percentage, so supply-chain inefficiency becomes something you can rank commodities by rather than describe.
- Filters that match how people ask. Year, division, product, season and price type — because the seasonal answer changes completely depending on which commodity you have selected, and the dashboard should make that discoverable rather than hide it behind an average.
The report
Built on the public WFP Food Price Database for Bangladesh, July 1998 – August 2024. Real screenshots of the working report.
1 Currency, distribution and inflation
2 Regional and seasonal pricing
Outcome
26 years · 14,178 observations
Four findings came out of it. Dhaka averages 928 BDT against Tangail’s 104 — close to a ninefold gap, driven by urban demand and longer supply chains. Palm oil carries the widest retail-to-wholesale margin of any commodity, which points at a specific supply chain rather than at inflation generally.
The distribution is heavily right-skewed — mean 204.48 against a median of 66.16 — so a small number of high-value commodities pull the average well away from what an ordinary household actually pays. Reporting the mean alone would misrepresent the everyday picture, which is why the median and mode sit beside it on the canvas.
What I’d do differently
The seasonal analysis is the weakest part. Combining all products gives Winter and Spring as the expensive seasons, but individual commodities disagree with each other and with the aggregate. The product slicer makes that discoverable, but the honest fix is a per-commodity seasonal index rather than one combined view.
I would also want to pair prices with a wage or income series. A price rise only means something against what people earn, and this dataset can only show one half of that.