Automated Python Code Generation with Mito

Key Takeaways

hello all my name is krishnak and welcome to my youtube channel so guys today in this particular video we are going to see how we can quickly analyze clean and manipulate the data using my to sheet now this miter sheet is an amazing library guys you'll be able to do very amazing things with respect to eda like exploratory data analysis and automated python generation code this entire video is basically done by jake who is from mito itself so please make sure that you watch this video till the end because as you all know that the first module in the life cycle of a data science project is eda right apart from that it also helps you to generate the python code automatically so definitely it is a win-win situation so definitely have a look on to this and was a video till the end so let's proceed hey this is jake from mido i'm going to give you a quick demo of how you can analyze clean and manipulate your data do some exploratory data analysis as well using mido if you don't know mito is a spreadsheet interface for python all you need to do is call on this interface and everything you do in this interface here is going to generate the equivalent python in the code cell below what i've done so far is just import this avocado data set and when i did that it automatically took this csv avocado file and turned it into a data frame for me it wrote this code for me i didn't have to write this code or any of the code you're going to see myself all i have to do is just interact in this interface here to get data into the sheet all we have to do is first call the sheet so you can import the mighty sheet package and call mightysheet.cheap which calls this interface here and once i've done that i can import data one of two ways so one i can pass in the data frame that i'm already working with has an argument here and then we'll see that data populate the sheet or i can go to the import model search my local files click the file i want and see that data imported here we've imported a data set we see one thing we noticed is that the csv file we had it had an index column already and we automatically automatically have an index in mida so i can delete this column just by clicking delete there we go and now i see my data so we have these dates and we have this information about this avocados that we're dealing with these different dates so i'm going to look at this date by month so we have these different months here so what i'm going to do is i'm going to add a column and then just really easily use this month function selecting this data and we see it will retrieve the month from each of these from each of these entries here in the date column and so now that we've done that we have this new column here i think i want to rename it just to make it a little simpler month here we see it updates and renames that column month and those two operations we just did that code is generated below here so automatically generates the code for us that we can use carried forward in our analysis or to automate a process or whatever we want to do with it now that i have these months let's say i want to look at the frequencies i have for the different months there's two ways i can do this two ways i can visualize this so i can select this column here and for any column we select by clicking on this button here i have three menus here so i can apply filters to that column so we have uh we have 18 000 entries here i can apply a filter let's say i just want the month of october so that would be 10. we get that we see now there's only 1500 so we can see how many entries from october we have and again when we apply that filter we get the equivalent code for that as well i'm going to take off this filter really quickly though we can also look at the values here so this is going to show us the different unique values we have in the column obviously here it's 1 through 12 and how many of those unique values we have for each of them and what percentage of the total amount of values that is so a really great way to see our values we also get the summary stats here we can sort of visualize this breakdown here we'll show you another visualization a little bit we get the mean standard deviation the quartiles max how many null values we have here is zero and lots of other good values as well and we get that for all the columns in our data set all you do is just click the button here i can also apply a pivot table to look at this data if i want to look at the counts of how many i have for each month so i'll click our pivot button here this creates a new data frame that we're creating with the pivot table here and we have the option to delete duplicate or rename any of these data frames that we have so different data frames appear as different sheets down here as my row i'm going to put month and then i'm gonna actually just do a simple value count of month here and we see for each month how many values we have so for one we have one nine four four et cetera et cetera and i can put a sort on any of these columns as well by going here and so i can put this in actually i want this in descending order and so we can see one january has the most most march is the second most et cetera et cetera june has the least amount of values i can also visualize this data as well using our graphing here actually and when i apply that sort you'll see the the code for it right there as well as the code for the pivot table down here and so now let's apply a graph we can show you some of our visualizations so i want to look at uh the graph i'm going to make is from df2 the chart type we can do bar chart box histogram or scatter plot let's do a bar chart here that's my x-axis i want the month very simply it's my y-axis i want month count and we'll see this chart here a little bit uh more detail than we had in the summary statistics box and we can look at this data we can hover over to see the values we can zoom in on specific values as well and if i want to copy the code from this visualization this is a really important feature because a lot of people when they're getting into python and even when they're really advanced with python getting the code read for visualizations can take a lot of time so with this feature here you can just generate your your visualization really simply with this menu here and then copy the code and we can paste it here and we have the code for this visualization in fact if i run this if i run this cell and then i would run this cell we get our visualization right here so you can use these visualizations carry forward in your analysis so let me close this and let's go back to our base data set so other operations we can do as i said add and delete columns we can export this as a csv we can undo and redo steps we also provide a step history here so you can see all the steps you've taken in your analysis and again all those steps are reflected in the code that is auto commented and i'll close this uh one other feature just to show you here is that we do have merging functionality so if i did want to join these two columns together i could do that i can decide what type of merge i want to do sorry not two columns these two data sets together i could do a lookup a left join a right join an inner join or an outer join we can decide what the merge key is going to be what columns we want to keep in the merge et cetera and that would as we see in df3 here generate the equivalent code there uh maybe that's the last thing is we actually do have a full screen mode so if you click that button you'll see the mighty sheet populate full screen and it will uh still generate code so mino's really powerful tool for generating python in this visual manner you don't have to leave your code you're not sacrificing code at all you still get all the same code all the same usable auto documented code but you get to do it in the visual environment so you don't have to go to stack overflow or google to look up syntax all the time it's really fast way to get your analysis done because that's what's important after all is finishing your analysis and having some good work done so i hope you check out mido again here's the install instructions and the full documentation and yeah hope you enjoy thank you