**The most surprising finding is that this dataset's "Amount" figures can't be taken at face value. The same company shows up in several months with very different amounts, and the biggest numbers cluster in April and are often formatted oddly.** **1. Pine Labs appears four times, and the April figure is far larger than the rest.** - January: $20M - February: $150M - March: $50M - April: $1.2 billion, with the stage marked "undisclosed" The other months' figures are small enough to read as separate rounds. April's is 8 to 60 times larger, which is hard to read as another round (step-1). **2. Other repeat companies show the same April jump.** | Company | Earlier figure | April figure | |---|---|---| | DealShare | $45M (Feb) | $624.8M | | Xpressbees | $300M (Feb) | $577.6M | | Livspace | $180M (Feb) | $431.6 (no unit, so it is unclear whether this is $431.6M) | | Chargebee | $250M (Feb) | $468.2M | | Mamaearth | $52M (Jan) | $110.8M | | Classplus | $70M (Mar) | $129.5M | The stage often stays the same, for example Series E for DealShare and Series F for Xpressbees. So the later figure looks like a restatement or a different measure (possibly a cumulative total), not a new round. The ratios range from about 1.9× to 14×, so it isn't a single currency conversion. I haven't confirmed the cause from the data. **3. April amounts use inconsistent number formats.** - Some use a period as the thousands separator, such as `$624.800,000`, `$110.800,000` and `$129.500,000`. - `$431.6` has no scale at all. - Because Amount is stored as text, a naive parse would silently turn these into tiny or wrong values. - Even `Securden` appears twice in April, written as `$11.700,000` and `$11,700,000`. **4. Missing and blank values are common.** - "Undisclosed" is the most common amount value (35 rows). Another 25 rows are blank (step-0). - In May, 19 rows are blank or misaligned (shown as `May::`, step-1). - A few rows are exact duplicates, such as NowPurchase's $2.4M seed in January listed twice. **Why this matters:** The card describes each row as one funding round, but these results suggest otherwise. Summing Amount across the five tables would double-count companies and badly overstate April, mostly through Pine Labs. Before any totals or trends are drawn, the amounts need cleaning (consistent separators, a de-duplicated company key) and a decision on how to treat the April restatements. Two follow-up queries I attempted this turn failed to parse, so I have not checked how widespread the April inflation is across all companies. The examples above come from the 14 repeat-company rows I did retrieve.