I've not listed the bottom 5 to save embarrassment (of course, there is nothing to be embarassed about) for the makers involved, but instead note that all products from both groups all still return successful HTTP responses. Ordering by votes we found the top and bottom 5 voted products of all time. Next, we take a look at products according against the votes they received. You can see how that breaks down in the chart below, where the split between response types is shown for all products. Over the entire dataset, I found 20,970 (or 22%) of products now produce some form of error. We consider a 2XX (Success) and 3XX (Redirection) status codes successful. Errors are categorized as one of 3 types based on their HTTP status code, or if another form of exception occurs (e.g. We used to have separate sections on the site for these types of products, which dramatically increased the number of "featured" posts because we essentially reduced scope." Ryan Hoover, founder of Product Hunt Types of ErrorsĪs described, we visited each product url to determine whether it still exists, or it now returns some sort of error. It's only more recently in the pandemic era from 2020-2021 that we see a marked increase in the regular number of products featured each month again.Įdit: "The drop is largely because we deprecated podcasts, games, and book categories. My thinking here is that other places to showcase products may have started cropping up around this time. Through to 2016, the number of products steadily increases before dramatically reducing in 2017. We can see how the total products break down to monthly featured products over the entire 8 year period we have data for. Our first query to the API is therefore as follows: In order to query any data, you also need to set an AUTH Bearer token. I'm using Python and HTTPX to make my queries, but you can use any language and library you're familiar with.įirst, we query the Product Hunt API to return pages of product details as JSON, ensuring to return as large a dataset as possible with the "per_page" param. We illustrate this with a simplified query of a single product below. Getting all the data necessary for this analysis was a 3 stage process. I've used their older API as it returns much more data per query (50 posts with all data vs 20 posts with limited data) enabling us to get the full dataset to work with in just a few hours. Product Hunt has two versions of their API currently available, one a standard REST based interface (deprecated) and v2 being based on GraphQL. I've queried it to return details of all their featured products over its history from December 2013 to the end of January 2022, which totals over 90,000 products. The source of all the data I talk about here is from the Product Hunt API. Along the way we'll also discover other interesting insights into the dataset. We collected information from all the featured products from Product Hunts 8-year history to determine how many of them still exist online or have disappeared into the tech wilderness. In this article I'll be showing how to query the Product Hunt API to collect data. But how many of the products of the past stood out from the crowd and continue to remain online today? Mountains of new products appear daily, complete with well produced intro videos with every new competitor bearing a striking resemblance to one other. Releasing any new product these days is a competitive business.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |