How often, having an array of texts (any subject) do we want to see thematic squeeze? What topics are being discussed? What are the most popular? How great it would be to enter search words and get trends around them!

The system can:
- Get an array of data and save it under your user.
- Build a list of topics with grouping documents by topic.
- To do realtime search in documents and build topics on the found selection.
All of these functions are available through the API, which is easy to use by connecting to one of the tariff plans (there is a free trial for a month!).
')
We indexed social. media (twitter, facebook, vkontakte) at the beginning of autumn 2015 and made screenshots of the system for visual assessment.
Consumer segment: Magnet.

Yota:

Politics: Putin's speech at the UN.

This is the topic from the inside:

The previous screenshot shows the content of the “UN 2015” theme, which includes 7 news items.
As can be seen from the screenshots, the system is a full-fledged search engine (with the support of Russian morphology) and is a convenient tool for quick navigation through arrays of information. We have no restrictions on the supported topics and data type - social. media, news feeds or your corporate documents. All described functions are available both as a SAAS product with a user interface, and as part of the Insider API, which can be embedded into your existing systems and mobile applications.
API principle:
1. Download data (posts, articles) to the system.
End-point: /articles/uploadJson : POST [ { "id": 2134657, "title": "", "description": " . . - .", "link": "https://twitter.com/palnom6/status/704979632127418369" } ]
In one request, you can send up to 50 texts. In response, the API responds with code 200, if everything went well.
Pay attention to the id parameter: by the value of this parameter we will get the binding of this post to a specific topic / trend in the next step.
2. Get trends: either by documents in general, or in conjunction with keywords.
End-point: /articles/cluster : POST { "query": "" }
In response, the API generates trends.
[ { "labels": [ "" ], "score": 1.510325122396045, "docs": [ "2134657", "2134656", "2134655", "2134654" ] }, { "labels": [ "" ], "score": 1.2447816860782057, "docs": [ "2134653", "2134652", "2134651", "2134650", "2134649", "2134648" ] } ]
The score value reflects the relevance of this subject to the user's request.
In step two, you can pass some keyword around which themes / trends will be built.
The Insider API is available here:
market.mashape.com/dmitrykey/insiderapiIn addition to trends, other convenient features are: