“Numbers don't speak until you put them in conversation with each other.”
Kolkata Cross-Domain Insights
Patterns that emerge when you connect data across 14 domains of Kolkata open data
Single datasets answer single questions. This page puts them in conversation — some 1,500 days of overlapping weather and air quality readings, alongside budget, mobility, health, housing and demographic series drawn from across the dashboard.
The strongest relationship in the data is also the least intuitive: temperature and air quality correlate at about −0.71. Kolkata's air is worst when the weather is most pleasant, because winter inversions trap pollution near the ground. Humidity, wind and rainfall all pull the same way — every coefficient is negative, in the −0.4 to −0.55 band. Cold, still days concentrate pollution; wet, windy ones clear it.
About this data
What this page shows
Seven collapsible cross-domain sections — climate and disease, mobility and safety, education and digital access, real estate and environment, healthcare, governance, and demographics — each pairing series from different sources with the correlation coefficients and charts behind the claim.
Where the data comes from
Every series on this page originates in one of the domain datasets documented elsewhere on this site: IMD and Open-Meteo weather, CAMS air quality, NCRB crime, KMC budgets, Census 2011, UDISE+, NFHS-5, TRAI, RBI HPI and others. Correlations are Pearson coefficients computed over the overlapping period of the two series involved.
Limitations worth knowing
Correlation is not causation, and on this page that warning is doing real work. Two series can move together because one drives the other, because a third factor drives both, or by coincidence — and with enough series compared, coincidences are guaranteed. The temperature–AQI relationship has a well-understood physical mechanism behind it; several others here do not, and are offered as patterns worth investigating rather than findings. Series also differ in granularity and vintage: daily weather compared against annual crime, or 2011 census data against 2024 telecom figures, cannot support fine-grained inference. Treat this page as a set of questions, not answers.
Embed this page
Drop these charts into your blog, brief, or report. Updates flow through automatically.