SPA FAQ

What is Smart Policy Assist (SPA)?
Smart Policy Assist (SPA), a product from Explore Digits, uses machine learning and natural language processing to analyze public comments or feedback to support effective policy making.

You can easily upload, analyze, and sort through an unlimited amount of public comment data.

Results are displayed in easy-to-use dashboards.

How does SPA improve the analysis of feedback for policy making?
By automating the analysis process, SPA provides faster and more accurate insights into patterns, trends, and sentiments in feedback data, compared to manual methods. For example, SPA identifies identical and similar comments, commenter’s organizations, and categories, and can be customized to support user’s workflows for reviewing comments.

How can SPA be integrated with other policy making processes?
The results from Smart Policy Assist can be integrated into policy making processes, as additional sources of information and evidence, and used to inform data-driven decision making. The power of NLP now allows policy makers to analyze unstructured data in the form of public comment and to respond effectively to those comments.

How can I get started with SPA?
SPA will be offered through Amazon Web Services (AWS) marketplace soon. Once available, SPA will need to be installed into a user’s AWS environment. However, if you would like to see what SPA has to offer, please click the “request a demo” button below or send us an email at spaadmin@exploredigits.com.

Do I need any additional hardware to use SPA?
This product is completely cloud-based and hosted by AWS, so you don’t have to worry about upgrading or maintaining software or servers. This gives you the flexibility to access SPA and run your analysis from anywhere with an internet connection.

What is data added into SPA called?
Data added into the application, regardless of the data source, is called an analysis.

What sources can be used to add data into an analysis?
You can add any comment data that is publicly available via Regulations.gov.
You can also add any comment data from a CSV file.

How long can I expect to wait for analysis results?
The time it takes to see analysis results depends on the number of comments added into the analysis.

Analyses containing 100 to 1,000 comments can finish within 2 hours of initializing analysis results.
Larger analyses containing 10,000 comments or more can take between six to eight hours.

How do the legal requirements for comment processing allow for use of SPA?
There are no conflicts between the legal requirements of manual comment processing and SPA.

Every comment is available for review by SPA users. SPA simply aids in the organization and categorization of the comments to make the review process more efficient.

Can I review more than one analysis at a time with SPA?
Yes, SPA allows you to import multiple analyses into the application.

How is SPA authenticated?
SPA utilizes AWS Cognito to authenticate users.

Is it possible to integrate the product with enterprise Identity Management (IDM)?
Yes! We can integrate into any enterprise IDM to meet your authentication needs.

Will all SPA users be able to add and review data in SPA?
SPA recognizes two different user roles within the system - reviewer and administrator. A reviewer can only review data that has already been loaded into the application. An administrator can review data and can also manage data in the application.

Can we personalize the front end to have organization themes?
SPA intends to add this feature in the near future whether as a feature through the UI or as a additional service the SPA team can offer.

What machine learning models does SPA utilize to enhance comment review?
Currently SPA offers two models: similarity and sentiment.

How does the similarity model identify comments that are similar?
The similarity model stores information on every single sentence appearing in each comment in an analysis. Each of the sentences are compared against one another.

When you review an analysis, unique comments are displayed by default, while similar comments are grouped and made available for review in a separate dashboard.

What does the sentiment model find within comments?
The sentiment model returns sentences containing language that our algorithms identify as having positive or negative connotations.