enform-banner.png
MIMS Final Project 2015

EnForm: Information Portal

http://enform.ddns.net

TL;DR

​EnForm is web search that offers a more contextual experience than your run-of-the-mill search engine.

  • Uses NLP to offer a profile of the full results corpus
  • Integration with open datasets drive rich visualizations alongside textual results
  • Innovative interface components optimized for rapid scanning of results
  • Tools to assemble content into narratives, preserving references

INTRODUCTION

We present a tool that pre-indexes text from the web and then employs a number of powerful javascript libraries to slice, dice and serve dimensions derived from them alongside reference tables of raw data collected from open APIs. Multiple evolutions of this platform have served as the more tangible of a larger set of ideas we had for how a research experience could be improved when accompanied by these contextual methods for interacting with a corpus.

Many search engines today serve as vehicles that send users off to various places in the vast world wide web; and user experiences vary greatly upon arrival. EnForm, or [En]ergy In[Form]ation Portal for short, is a search engine that withholds that fundamental ability (find and go), but also offers a different approach. By maintaining a collection of scraped web resources and pre-processing them with NLP techniques, such as NER (Named Entity Recognition), we can extract a number of new dimensions from the corpus of the results and use them to offer new perspectives on search results.

Searching the web for answers to complex issues is a cumbersome process. Modern search interfaces provide little additional context about the topic in question, leaving the burden to the user to create their own retrieval processes. Assembling information is therefore divorced from the search experience, resulting in a highly fragmented and inefficient experience.

We envisioned a platform that collects, synthesizes and visualizes structured and unstructured data sources to support integrated research and analysis. Ultimately we aim to provide an enhanced contextual tool set to improve both the efficiency and thoroughness of answering a question.

Any question? Well yes, ideally. But in the span of one semester we realize that scope must be limited. Which brings us to the question we've gotten most...

Why Energy

In the energy industry, the range of data sources, types and disjointed access to them is especially acute. Two of our team members have strong ties to the energy industry. These were the main motivators for the decision to limit the initial collection to energy-related resources. However, EnForm was built as a content-agnostic system, meaning that if we were to flush the database and populate it with resources about anything else, such as textual assets from the library of congress, and quantitative data about congressional voting history, it would work just the same.


By giving users the tools to flexibly navigate, sort and filter their information, we hope to help people uncover relevant data more quickly and simplify the task of collecting and sharing findings as a cohesive story. 

Last updated: October 7, 2016