Arbisoft partnered with AEE to help create a future of secure, clean, and affordable energy by making informed decisions based on data. Our team of data experts crawled, optimised and automated large amounts of data from important policy websites.Tell us about your project
“We now have more data than the largest online encyclopedia as a result of leveraging Arbisoft’s expertise. They’re definitely web crawling experts.”Eric FitzVice President, Engineering and Product Development
Advanced Energy Economy or AEE is a group of businesses on a mission to nurture a prosperous economy powered by secure, sustainable, clean and affordable energy. They do this using policy advocacy, analysis and education — and to do that they need access to data. A lot of data. AEE’s product, PowerSuite, collects vast amounts of public data from Public Utility Commission websites for each U.S. state and then provides it to PowerSuite subscribers in an easy-to-understand aggregate they can use to obtain updated information, track long-term trends, and most importantly make better decisions. And this is where Arbisoft comes in.
To be competitive, PowerSuite required daily data crawls of 50 large PUC websites. This entailed massive overhead costs in time, performance, and API service expenses. There had to be a more efficient way of doing this–and they came to Arbisoft to find it.
We implemented a custom-designed distributed crawling mechanism that used automated intelligent data extractors capable of semantic analysis. We optimized their system and added the ability to clean up data in realtime, further cutting down time between data collection and data usability. We also provided a REST API that fed data to their user-facing Ruby on Rails app. The high performance API was initially implemented using Google App Engine and later transitioned to the Django REST Framework app.
Our work made data fetches hyper efficient, reducing crawler runtime from over a week (168+ hours) to less than 4 hours. This represented a whopping decrease of 97.6% in minimum runtime duration, significantly enhancing the ccrawlers’ ability to capture and present new information to subscribers as quickly as possible.
Despite the massive increase in performance, the expense of data fetches also reduced to half of that incurred by the conventional method. Shifting the API form Google App Engine to the Django REST Framework also yielded an additional $600 in savings per month for an approximate $7200 in savings per year.
Python, Django, Scrapy