HP Autonomy’s IDOL OnDemand APIs Nurture Apps Ecosystem

by Darryl K. Taft

With free access to a “try and see” explorer available at IDOLonDemand.com, developers have access to a growing catalog of RESTful APIs based on HP IDOL, including format conversion, text analysis, text search and image analysis. The IDOL OnDemand platform enables rapid prototyping and app development, engagement and feedback via the IDOL OnDemand Developer Relations Team and collaborative support from an online community.

This proof of concept application using the IDOL OnDemand Face Recognition API lets a Google Glass user take a photo of a person for image analysis by IDOL OnDemand. Once the photo is taken using Google Glass, the app sends the photo to the IDOL OnDemand Web services, where IDOL OnDemand analyzes and matches the face in an image database, confirms the name of the person for the matched face and then displays the name of the person in Glass.

This demo app combines the IDOL OnDemand OCR and Query APIs and turns business cards into Salesforce leads with minimal user input. The business traveler opens the app and signs into his or her Salesforce.com account. After simply taking a photo of the business card, IDOL OnDemand analyzes the image with OCR to extract the text. The user is able to edit the text as needed, and then saves the lead to Salesforce.

The Intelli’Stocker iOS mobile app, developed during an online challenge, uses the IDOL OnDemand Sentiment Analysis API to determine sentiment around stocks from a broad content feed from online sources like Facebook, Yahoo News and numerous RSS feeds. The app provides a simple dashboard that shows the general sentiment around a stock in real time. This helps users get an immediate feel for the positive and negative perception of a stock, helping them make more informed investment decisions.

Expire Social helps people manage the content they share across social channels. The mobile app recommends how long to leave a post in place, based on its content and then automatically takes it down for the user after the assigned time period expires. The app uses IDOL OnDemand technology to automatically profile both text and image content across the user’s social footprint.

The SFDC Case Index and Search demo app powered by IDOL OnDemand lets Salesforce.com users index and search unstructured, untagged information, including case chatter, emails, comments and attachments from within the Salesforce environment. IDOL OnDemand automatically provides the case sentiment (positive/negative/neutral), a list of similar cases, and a visualization of related concepts from all the other related cases so that users can quickly find information no matter where it is hiding.

HP’s largest show of the year, HP Discover, is using IDOL OnDemand to help conference attendees find sessions of interest, based on sessions they are already slated to attend. Making large events easier to navigate, the app uses the IDOL OnDemand Create Index, Add to Text Index, Query Text Index and Find Related Concepts API’s to identify similar events to the ones that attendees research and attend via the Discover Website.

This proof of concept app created in a hackathon uses IDOL OnDemand APIs to count the occurrences of specified logo and brand names in a video, providing brands with insights into how their logos and names are being used, and if they are being represented properly and in an authorized manner.

The Informme proof-of-concept app was created in a TopCoder challenge and uses IDOL OnDemand Language Identification and Sentiment Analysis APIs to analyze Wufoo online form data and automatically extract sentiment from those submissions in real-time. With this information, Wufoo survey administrators obtain valuable insights into what customers think of the user’s product or service, even about topics the user may not have considered. An optional feature lets users receive email notifications when Informme has received a submission and analyzed it for its sentiment.

ImprovedIn is a proof-of-concept Web app, created during a hackathon, that uses IDOL OnDemand APIs to analyze LinkedIn profiles and suggest recommendations to improve them. For LinkedIn users, the ImprovedIn app uses the IDOL OnDemand Face Detection, Highlight and Sentiment Analysis APIs to conduct a LinkedIn profile analysis. The resulting personalized report recommends content changes that will make the LinkedIn profile more effective and stand out from other profiles.

These browser plug-ins demonstrate the simplicity and ease of adding IDOL OnDemand API functionality to apps to allow users to perform rapid analysis of human information with a single click. Three apps were developed in an internal hackathon: A file viewer plug-in that lets users mouse over a ZIP file online to view the contents list without having to download or open anything; a Twitter sentiment plug-in that lets users execute a sentiment analysis and visualization of a Twitter stream right in the browser; and a “related news” plug-in that shows the user news articles that are related to any text he or she highlights in a browser from any Web page.

PwC built an app that provides analytics on social data using IDOL OnDemand, Hadoop and Vertica. The app performs sentiment analysis on external data such as social media and unstructured information like product discussions and consumer feedback. Project lead Dhiraj Pathak said, “Using the easy-to-connect APIs and not worrying about the underlying infrastructure created an enormous amount of acceleration for my project. HP IDOL OnDemand is creating opportunities, expanding footprints and breaking new ground in the cloud.”

HP’s Personal & Printing Systems (PPS) group used HP IDOL OnDemand APIs to identify duplicates from more than 110,000 documents in its software/driver repository. The group was able to identify that as many as 75 percent of documents were duplicates and candidates for elimination. The avoidance in manual labor required to review and migrate all documents resulted in an estimated cost savings of $150,000.