4 new time-saving tricks in pandas

In the last 20 months, the pandas library has been updated 10 times, introducing hundreds of new features, bug fixes, and API changes. In this video, I'll show you 4 new pandas tricks that will make your life easier!

1:18 Create a datetime column from a DataFrame
4:24 Create a category column during file reading
7:45 Convert the data type of multiple columns at once
9:48 Apply multiple aggregations on a Series or DataFrame
13:14 Bonus: Download the official pandas cheat sheet

== RELATED VIDEOS ==

== OTHER RESOURCES ==

== JOIN DATA SCHOOL ==

[jalego]

The last one is really really awesome! Thank you, Kevin!

[jaquesderasmo]

Thank you Kevin, can't describe how your videos are so amazing and helpful not matter who new or old they are.

[Nintenzo]

At 9:30, tried using inplace=True instead of assignment. No warning, yet the drinks' columns' dtypes were still int64 instead of float64.
drinks.astype({'beer_servings':'float', 'spirit_servings':'float'}, inplace=True)
ver. 0.24.2

[kostasnikoloutsos]

It's been a long time see you again Kevin. I am happy to see you again!

[liningwang]

welcome back to here, Love to see your video!

[arrahman]

Thank you Kevin, your videos are amazing. i got confident and interested to learn pandas and practice it.

[TobiasWeg]

Thanks for your awesome videos. For me it is always the go to for the 'How was that again?'
moments;)

[The_Bri_Cli]

Hello Kevin. Love the video series. I am wondering, many videos are microscopic views of pandas. I am wondering if you can do a video series that is more macroscopic. For example: 1) How to read multiple csv files from one folder and iterate over them some function/program and then shooting it out. 2) Common problems to debug in pandas like when a csv has characters that pandas can't read. These big picture problems I find to be the most troublesome.

[Maik.iptoux]

Hey looking your videos to find solution for my problem. I have SQL data in pandas dataframe (IP's +Hashes) now I want an count how often an hash is in DB as new column, also an list of all IP's in this hash. My problem is, that the count of IPs always overwrites the count of Hashes... I dont find any solution for a week now.

[durowbird]

Thank you for your video, very happy to see you again.

[mselim-op]

fantastic! Thank you very much indeed for your time!

[musafaugurlu]

Thank you, Kevin. You are great! But honestly, I like your other style which you are writing codes during your videos. it is a lot more helpful.

[ankushsharma-guco]

Thanks a ton! Your videos have helped me a lot.

Thanks for your content, amazing as always.

[chiragpalan]

can you please make couple of vedios of real world data analysis problem using pandas?

[joeyzhou]

how to convert data type of object to int? Do not know why my data shows numbers as object data type. Thank you!

[rajeshd]

Hi Kevin,
Thanks for the wonderful vedio... I


have question regarding your Tip 3. i have tried below code snippet; but unfortunately Column Continent is not converting to Category data type. why there is strange behavior. it is converting as expected in case of beer_servings to Float


drinks.astype({'beer_servings':'float'}, {'continent':'category'}).dtypes


Output:
country object
beer_servings float64
spirit_servings int64
wine_servings int64
total_litres_of_pure_alcohol float64
continent object
dtype: object

[alextj]

Is there a systematic way for pandas to determine which dataframe column may be helpful to convert into categories, or will that generally be contextual?

[opalkabert]

Your last trick is cool but talking about df.describe() I have used pandas-profiling package and it is awesome and gives you more insight. Good for EDA.
Thanks for sharing these tricks

[zuolingxiao]

Welcome back! finally got some new videos to feed myself :P