AutoPlotter - A GUI based Exploratory Data Analysis in Python

Key Takeaways

exploratory data analysis is one of the most important part of the data science process because it allows you to gain an understanding of your data set and such preliminary understanding will allow you to plan your own data science projects as well as shaping the direction at which the project is heading for and so in this video we're going to be taking a look at the auto plotter python library which will allow you to perform a point-and-click or gui based exploratory data analysis of your own data set and without further ado we're starting right now so this illustration was drawn by myself in collaboration with kenji and so you could check out this particular article and also the video described in this article as well and so i'll provide that in the video description and so let's head over to the auto plotter on the pi pi as you can see here it will allow you to do a do i based exploratory data analysis and this report that will be generated by auto plotter library is built on top of dash and dash is a python library from the same creator of plotly which allows you to build an intuitive and interactive web application to explore your data sets so let me know in the comments if you're interested in a tutorial video of dash so let's head over to the google code lab that i've created and the first thing that you want to do is you want to install the auto plotter library and in order to do that you could type in pip install auto plotter and let's do that all right and it's finished and now let's import the libraries and we're gonna import the run app from auto plotter and then we're also going to import pandas as pd and now we're going to run it all right and we're going to import the nba player stats data set that was mentioned in one of the prior videos on my youtube channel and you could get the data set from the github of data professor and so let's have a look at it as you can see it is the data set of the national basketball association or nba and it has all of the player stats data and it was obtained from the basketballreference.com website and i'll provide you the link to the video describing the web scraping of this particular data set let's head over back to the notebook and we're going to run the cell so that we will assign it to the df variable and we're using the pandas read csv function and let's have a look at the data frame that we have read in and as you can see here it is comprised of 708 rows 29 columns and you can see that we could simply run the app by using the run app function which we have imported prior to this and the input argument here will be df which is the data set and then you could decide on whether you want to run it in either the mode of inline or external mode and so inline will allow you to run it directly from within the google code lab or a jupyter notebook and if you and if you run it in an external mode you could run it in a separate window of its own so let me show you both of them let's start with the inline and it's loading and now you can see the data frame and notice that you could navigate through each of the page and each page will be comprised of approximately let me count 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 15 rows and you could scroll the mouse to the left and right and you could click on the sorting icon here in order to sort the data and you could also filter the data by typing in the operators like for example if you would like data greater than 0.5 just type greater than 0.5 hit on enter and now the data will only be those having values of greater than 0.5 and now you see that the number of pages reduced to 23. let's try less than 0.5 let's try less than 0.1 okay so you see that the total pages has decreased over here have okay this one works shooting guard center or how about the team t-o-t l-a-l lakers okay so the team works the position works and how about age let's say that we want players greater than 30 years old we could type in greater than 30 and we see that there is approximately eight pages here and how about less than 30. so we can see that most of the players in the nba were aged less than 30. so there's 38 pages in comparison to eight pages here all right so this is for quickly exploring the data sets in the data frame and this is right here in the data preview icon and let's head over to the plots icon here click on it and you can see that you could use either gc plot 2 seaborn plotly plotly dark or also the presentation version and please note that the presentation version is essentially in white background and let's give heat map a try and you could select x-axis y-axis data let's go with the team and then y-axis let's go with the position and the value let's select the age and then here you're going to see that the age ranges from 20 to approximately 35. you can see according to the color for each of the positions here against the team here you can see the relative age of the players so if it's yellow then the player is older and if it's purple then the player is younger so you can see the relative age between the positions and the team and let's choose another parameter here let's try the three point percentage and here you're going to see that atlanta has a shooting guard which has pretty high three point percentage and another team would be milwaukee and indiana for power forward and shooting guards choose another one about two point percentage all right and now you're gonna see that evenly distributed throughout teams in the nba and so you could also try out other types of graphs as well and let's head over to the analysis part click on the icon here and the first tab here is data distribution and so here it is selecting player and let me deselect it and let's select the team and then you're gonna see the histogram for each team you're gonna see the count of the occurrence here and let's try another let's select another one let's say age and then scroll down you're gonna see the histogram of the age distribution let's try another let's try three point percent and then you're gonna see three point percent here so there are zero percent and then in the range of about 0.1 until approximately 0.524 where you could see that there is a greater portion at approximately 0.275 until 3 7 all right and on top here you can see the box plot and let's see let's click on the distribution and then you're going to see let's wait for it all right for the distribution of h you're going to see it as a line here same for the distribution of three points okay because this is categorical data let's click on the statistical analysis and so you're gonna see the descriptive statistics for the count mean standard deviation minimum value maximum value 25 percent 50 and 75 quartile for each of the columns here let's click on the association tab and then you're gonna be able to compute the correlation between let's say age and the games played and there's a pretty low correlation so there's no correlation there let's try age and about three point percent no correlation there about free throw is there any correlation not significant but slightly higher how about free throw and three point somewhat higher correlation okay so let's try the external version run it and then you're going to notice that it creates this url click on it and from within collab it will launch this new window and then you could also perform the similar analysis that i've mentioned directly in here okay and so if you're finding value in this video please give it a thumbs up subscribe if you haven't yet already make sure to hit on the notification bell so that you will be notified of the next video and as always the best way to learn data science is to do data science please enjoy the journey