Photo by Evan Habeeb, USA TODAY Sports
In part five of the beginner series, we learn how to rank players in our DataFrame using the rank DataFrame method. We rank players based off utilization and Fantasy Football output and try to find regression candidates for your 2020 draft.
If you have any questions about the code here, feel free to reach out to me on Twitter or on Reddit.
If you like Fantasy Football and have an interest in learning how to code, check out our Ultimate Guide on Learning Python with Fantasy Football Online Course. Here is a link to purchase for 15% off. The course includes 15 chapters of material, 14 hours of video, hundreds of data sets, lifetime updates, and a Slack channel invite to join the Fantasy Football with Python community.
In this part of the beginner series, we are going to begin to look at our DataFrame we created last time a bit more closely and try to find some regression candidates for the 2020 season. If you have not already read part one of the series, here's a link for that. Part one of the series starts you off at the absolute basics of Python. If you want to instead skip ahead to where we start using Pandas, check out part 3 instead.
We are first going to do this using two Pandas DataFrame methods known as rank and sort_values. Remember the couple last times we started going over some of our first DataFrame methods/attributes such as head and columns. rank and sort_values are both methods and behave similarly to the head method we've used before.
Let's start by bringing in the DataFrame we created last time. Either open the Google Colab Notebook you were working with last time, or open a new Google Colab notebook to regenerate our DataFrame back to it's current state.
You should get the output below. If you do not already have access to the 2019 csv file we are importing in to our code here, here's a download link for that.
Remember that your CSV file needs to be uploaded to be able to run this code. You may need to re-upload it. In part 4, I give instructions on how to do that.
This is all of the code for this part, and although like last time it does not seem like a lot, it is actually quite dense if you haven't coded before. We'll go over the output and the DataFrame we get back in just a minute.
In that first line, we are creating a new column based off another column (Usage) and finding each player's Usage rank. This method also takes a keyword argument. A keyword argument is an argument that you have to explicitly define in the functions parameters. If you look ahead to the third line of our code, you'll notice that head doesn't require a keyword argument. Some arguments in a function have to explicitly specficied and some arguments can be assumed. We'll get in to the reasons for this in future parts, but it boils down to the fact that functions can have optional arguments. You can find our which arguments go in which functions by reading through the documentation.
In the case of rank, one such argument it takes is ascending. If we specify that ascending is equal to False, we are telling Pandas that we want to rank the largest numbers in the column we are ranking higher. CMC had the higher usage numbers and so he gets a 1 for UsageRank
We repeat the same process for FantasyPoints. We repeat this process becuase we want to find gaps between Usage and Fantasy Points scored on the season.
Finally, we use another method I talked about called sort_values. This method sorts our DataFrame by a certain column. By default, it sorts the columns by ascending order (smallest to largest). Since we are sorting via the rank column, this is exactly what we want. It takes in a keyword argument of by, so we can tell Pandas which column we'd like to sort on. (It also takes in an argument of ascending, in case you wanted to sort your DataFrame in descending order of a column. In this case, we don't.)
We then take on head to the end of that last line, because a DataFrame sorted via sort_values is still a DataFrame. We haven't discussed this previously, but head also takes in arguments. By default, head gets you back the top 5 rows of a DataFrame (By the way, try to figure out what running the .tail() method does), but if you specify an argument, that argument will override it's default of 5. Here, we specify 15, which means get us back the top 15 rows in our DataFrame.
Here is the output from the code we ran above. In the next part, we'll be examining these "gaps" between Usage and Fantasy Points further and trying to figure out what it means for your 2020 draft. We'll start talking about the relationship between Usage and Fantasy Points scored and what it means when a player has a low Usage ranking but high Fantasy output. For now, try to examine the DataFrame and find some gaps between Usage and Fantasy Points. Two interesting cases that pop out to me are Aaron Jones and David Montgomery. Montgomery had the 12th highest utilization amongst RBs but finished RB24. On the flip side, Aaron Jones finished 10th in utilization in RBs but finished RB2. Could this indicate that Aaron Jones is due for a negative regression to the mean and Montgomery is due for a positive regression to the mean? We'll find out in the next part of this series.
Thanks for reading, you guys are awesome!
