Looking at Product Hunt API data for the last 6 months

#Looking-at-Product-Hunt-API-data-for-the-last-6-months

Since we are launching Kyso on Product Hunt (PH), it would be pretty fun to showcase what Kyso can do by looking at PH's own publically available data.

So I spun up a Jupyterlab environment on Kyso and got to work.

This notebook only handles the plotting of the data, but the attached fetch.ipynb notebook shows how I downloaded the data, which you can access in the files tab.

I downloaded the posts for each day for the last 6 months and how many votes they recieved.

Loading output library...
Loading output library...
Loading output library...

Posts per day for the last 6 months

#Posts-per-day-for-the-last-6-months
Loading output library...

So here we can see that on an average day about 100 posts are submitted, but there is a lot less activity on the weekend, lets plot the total posts per day of the week.

Loading output library...

Yup - a lot less activity on Friday and the weekend, probably because people are away from their desks. Let's look at the votes.

Votes per days for last 6 months

#Votes-per-days-for-last-6-months

Please note these are only the votes made on the day of posting, as that is what a product's ranking on any given day.

Loading output library...
Loading output library...

No suprises here, same pattern, about 2000 votes made on a typical day and less votes on the weekend.

Distribution of votes

#Distribution-of-votes

This is the most interesting graph - on the x axis we how many vote a post got, and the y is how many posts how that number of votes. I've cut the y axis at 50 because over 600 posts only got 2 votes, so it dwarfs the rest of the chart.

Loading output library...