## The most surprising finding: a few dozen deals account for most of the money, and the largest one is probably a data error **1. A tiny share of deals holds most of the dollars.** - Of the 3,044 rows, 2,065 have a parseable USD amount. They total about **$38.06B**, but the **median deal is only $1.7M**. - **73 deals of $100M or more** (about 3.5% of valued deals) make up **63.6% of all dollars** (query step-10). - The typical startup deal is a small cheque. The headline "total funding" figure is mostly a handful of mega-rounds, so a mean or sum would badly mislead anyone benchmarking a raise. **2. The single largest record looks implausible.** - The top row is **Rapido Bike Taxi, Series B, 27/08/2019, $3.9 billion**. That is **10.2% of all dollars in the dataset** (`top_share` ≈ 0.102, step-1). - It is more than 1.5× the next largest deal, Flipkart's $2.5B private equity round in 2017. Flipkart and Paytm, with rounds of $1.0B–$2.5B, are the kind of companies that normally hold the top spots. - A Series B at that size, against a $1.7M median, strongly suggests a scale or entry error. That is my inference; I haven't verified it against an outside source. If it is wrong, the true total and the concentration shares shift materially. **3. The data has other traps that break naive aggregation.** - One row, **Drums Food (21/07/2016)**, has a non-finite amount. It sorts to the top and turns any plain `SUM`, `AVG` or `MAX` into `nan` (steps 0–9). Excluding it gives the clean totals above (step-10). - **978 rows (about 32%)** have no usable amount at all, so any dollar-based ranking covers only about two-thirds of the deals. **Bottom line:** funding in this dataset is extremely top-heavy. Use medians and exclude or verify the outliers, starting with Rapido's $3.9B and the Drums Food row, before drawing conclusions about vertical, city or round-type "typical" ticket sizes.