Trends over time
Track any indicator across selected districts, year by year.
7-year time series
/
/
Line chart shows the selected variable across all seven years for each selected district. Use the districts dropdown above to add or remove districts from the comparison.
Geographic distribution
District-level values for a single year, mapped across Bangladesh.
64 districts
—
Lowest value
—
Highest value
—
District with highest value
—
National mean
Value scale
lowhigh
Top 8 districts — selected year
Circle size and colour both encode the selected variable's value for the chosen year (teal = low, gold = mid, rust = high). Hover a district for exact figures.
Bivariate correlation
Explore the relationship between any two variables, district-by-district.
Pearson r
—
Pearson correlation (r)
—
R² (variance explained)
—
Interpretation
—
District-year points
Each point is one district in one year. The gold line is the ordinary-least-squares trend. Correlation does not imply causation — use this to generate hypotheses, not conclusions.
Multiple linear regression
Fit an OLS model with several predictors at once and inspect coefficients, significance and fit.
Ordinary least squares
Choose an outcome and at least two predictors, then run the regression.
Gradient-boosted model & SHAP
An XGBoost-style boosted tree ensemble, trained in-browser, explained with Shapley-value attributions.
GBM · exact Shapley
80 / 20
Random split
70 / 30
Random split
60 / 40
Random split
Time-based
Train 2017–2021 → Test 2022–2023
Pick a target, keep the default features (or adjust them), choose a train/test split, then train the model.