Comparing values in Pandas with "diff" and "pct_change"

Want to know how much the values in a series or data frame have changed from row to row? Meet the "diff" method, which will calculate this for you -- either from the previous row, or from any other row. And if you want to know the percentage change? Just use "pct_change" instead.

[@AndyWallWasWeak]

off topic Q, but inspired by a moment from this vid. When writing own functions, should I write for pd.Series or 1-column DataFrame? Sounds like could be multiple things to keep in mind when deciding

[@ReuvenLerner]

I'd suggest not writing anything for a 1-column data frame. Either write for a data frame (regardless of size), or for a series / column. I think that the latter is probably a better way to go, overall.

[@tyl9680]

What about diff by different categories? Say I have corn, rice, beans and wheat prices in the same df, and I want to compare the price changes within the same catogories.

[@ReuvenLerner]

You can totally do this! Just use "diff" on the result of a "groupby". For example: df = DataFrame({'category': ['wheat', 'corn', 'rice', 'wheat', 'corn', 'rice', 'wheat', 'corn', 'rice'], 'price': [10, 8, 6, 11, 7, 5, 15, 9, 4]}) df.groupby('category')[['price']].diff() You'll get a new data frame back (thanks to the double square brackets around 'price'), showing the difference for each row from the previous occurrence of that category. However, if you want to know which category is which, you'll probably want to join it back to the original data frame: df.groupby('category')[['price']].diff().join(df, rsuffix='_df')