Project Cloud Conversations

Project Cloud Conversations, a TIBCO LABS™ project, is focused on offering new ways of exploring and interacting with enterprise data.  Using natural language processing (NLP), machine learning (ML) and the now ubiquitous paradigm of a 'chatbot' user interface, Project Cloud Conversations’ goal is to provide enterprise users with a simple-to-use tool for querying enterprise knowledge.

Provider

TIBCO LABS™

License

None

Overview

Project Cloud Conversations by TIBCO LABS™ provides capabilities for users to query knowledge using natural language.  The current Project Cloud Conversations' underlying data sample (talk to us about bringing your own data!) is based on a knowledge domain built from public data provided by DonorsChoose.org.  DonorsChoose.org data represents donations made to public schools in the US in 2017.  Project Cloud Conversations leverages the power of a graph database by mapping enterprise data to a knowledge base represented as entities and relationships.  The user interface is as simple as a typed chat; you can enter queries in natural language.  As an example:

"show me the donations received in q1 2017 per month"

Project Cloud Conversations can find answers to questions that were not anticipated and produce meaningful responses in the form of textual information or graphics.  Project Cloud Conversations makes it easy for you to have a “conversation” with your enterprise data!

Submit Feedback

Release(s)

Project Cloud Conversations Release

Published: September 2018

Quick Start here

There are currently no reviews for this content.

Why not be the first to review it - click here to login

Project Cloud Conversations

Overview

Project Cloud Conversations by TIBCO LABS™ provides capabilities for users to query knowledge using natural language.  The current Project Cloud Conversations' underlying data sample (talk to us about bringing your own data!) is based on a knowledge domain built from public data provided by DonorsChoose.org.  DonorsChoose.org data represents donations made to public schools in the US in 2017.  Project Cloud Conversations leverages the power of a graph database by mapping enterprise data to a knowledge base represented as entities and relationships.  The user interface is as simple as a typed chat; you can enter queries in natural language.  As an example:

"show me the donations received in q1 2017 per month"

Project Cloud Conversations can find answers to questions that were not anticipated and produce meaningful responses in the form of textual information or graphics.  Project Cloud Conversations makes it easy for you to have a “conversation” with your enterprise data!

About the Knowledge Domain "donors"

Project Cloud Conversations applies the notion of knowledge domains ("groups of data") to the data you want to query.  You could have various domains loaded concurrently into the system, including sales data, product and marketing data, customer data, etc.  For this initial project, we have pre-loaded the sample domain "donors" based on public data from DonorsChoose.Org.  It is an interesting data set with more than a million donations.  The knowledge domain is stored in a graph database in the form of entities (the basic notions) and relations.

The notions in this domain are :

schools have projects
schools are located in a state
donors provide donations to projects
projects have a list of resources, each resource concerns a item and a vendor

From the Project Cloud Conversations chat interface, type 'help' to get the list of entities and more information about each one.

Basic queries

You may start by a simple request such as:

"show me the list of projects"

You will see that the bot will reply with a list of "cards". The bot limits the response to a manageable count of 25.  You can browse the list and click on an individual card to get more detail.

When you access a detailed card, basic information is displayed (the main attributes or "predicates" of the selected entity).  You can expand the card information viewed by clicking the "..." at the bottom of the card.  Some information displayed on the card is about 'relations' to other notions.  For example, the project card will display the school and also the count of donations for this project.  You can click on those links to 'navigate' through the data.

There is usually not a single way to ask a question.  For example, all of the following will produce the same result:

"show me the projects"
"could you show the list of projects"
"list of projects"
"projects"

Filtering

You can add criteria to your request to filter data.  For example :

"schools in state california" 

When you receive the result, note that more than 5000 schools were returned.  You may want to narrow down your search.  Do so by 'continuing the conversation'.  For example simply type:

"with at least 100 projects"

The bot retains the context of your conversation and applies the new criteria to the current list!

Counting

The bot is good at counting things:

"how many donations are there"
"count of projects in state idaho"

Counting per category (aggregating)

"Help" will tell you how different notions are related.  You can count things by specific relations.  For example:

"count of projects per type for state california"

While typing you will notice that the bot is trying to help you in building valid queries, such as when you type 'per', a list of valid relations will appear below the chat box.  For example, "schools per state" will count by category, and since state is a geographical notion, the bot will decide to display the result as a map.  In each case you can click on the graph elements to get a list for a specific category.

Computing

The bot has basic computng capabilities, including sum, min, max. average.  For example:

"average donation in idaho"

Dates

Dates are related to the notion of "what can happen" on an entity.  For example, a project is posted, and a donation is received.

You can ask:

"projects posted in q1 2017"

You can also aggregate the information using day / month / year hierachy:

"projects posted in q1 2017 per month"

"Try Now" and Feedback

Start exploring Project Cloud Conversations today and discover what you can do with it.  Click the "Try Now" button above or here.

Additionally, we would love to get your feedback on your ideas, including:

- applicable business use cases
- what additional features would you like to see
- what do you like
- what you don't like, or found frustrating
etc.

All feedback / comments / questions are welcomed!

Other Links

View the Wiki Page