Analytics & Visualization
Table of Contents
Visualizing Complex Data
Our analytical and visualization tools help data scientists visualize and decipher complex data in an intuitive manner. The tools are scalable & can be deployed as enterprise analytical and delivery platforms.
Explore
- Start with preferences
- Explore via:
- Topics
- Content types
- Concepts
- Content metadata
Analyse
- Top actors in the notifications pool
- Most common interactions amongst actors
Spot trends
- Profiling
- Generic trends
- Gap analysis
Parallels
Case Study
Problem Statement
- Solution for a Personalized research content alert and analytics system
- Given a pool of diverse content, both ready to consume and to be acquired from open-source periodically, how do we alert users on the most relevant content in the subject(s)/area(s) of their choosing? How can we make sure we cover most of the latest research?
- How do we make sure we cater to specific user preferences without explicit input(s) or overpowering users with heavy volumes?
- We need a solution that is sustainable; robust in terms of core capabilities and flexible in terms of content and functionalities
Challenges
- User Queries
Users are allowed to enter almost free text selection of keywords/phrases with minimal syntax.Onus is on the solution envisioned to incorporate a query parsing component
- Content
Multidimensional content and metadata to be leveraged across sources with due diligence to update cycles
- Preference Management
Completely automated preference tracker and a feedback inclusion system to act on and update the system to tailor the alerts to the preference pattern with immediate effect
- Workflow
A adaptive workflow system to ingest incoming user subscriptions, unsubscribe requests, mail tracker and change management
- Analytics and Decision Aids
Analytics and decision aids that point the decision makers towards how well their content is being accepted, consumed and used. This system ties in content, user behaviour and approaches used for suggesting relevant alerts
Solutions
Solution architecture includes thorough and robust systems for;
- Content ingestion : High Throughput parsers/spiders for content acquisition, transformation and standardization
- Semantic enrichment : Machine Learning models and taggers to semantically enrich the metadata with relevant concepts, making the data discoverable, reusable and traceable
- Query parser : A natural language query parser to identify concepts, contexts and intentions from user keyword(s) and translate it to a query language that can be used to retrieve content from data lake(s)
- Mail scheduler and user management parallels : A secure, fast and traceable system that can be configured to set up and manage actual alerts to subscribers
- Analytics and Decision aids : Integrated usage reports and infographics to assess the performance of the system, both qualitatively and quantitatively
Together these five components, heavily backed by AI modules make for an excellent, up-to-date, scalable and quick to integrate and feedback inclusive solution for the
What gives this solution an edge?
- Plug and Play content acquisition systems
- Semantic enrichment with no prior/custom requirement and a provision to plug in custom ontologies/CVs if necessary
- Scalability – Around 3500 users per week*
- Intelligent search and retrieval with due diligence to personalization – With little to no additional inputs from users
- Integrated workflow components allowing decision support and feedback ingestion
- Generic core modules that can be easily integrated with existing systems and/or adapted for a wide range of other applications
