How to sort, tag and cluster your data in ?
3 easy steps
Upload XLSX file
TopicMap allows you to get a quick overview of the texts without actually reading them. All you need to do is to pack them in an excel (xlsx) file and upload it (click on "Create Insights"). The TopicMap is a result of a process called "Topic Modeling". A smart algoritm bootstraps some topics represented by keywords and adjusts them iteratively so that they reflect the documents in the most complete manner. The result is represented with a graphical topic map and an excel where each text is assigned just one topic.
You can also use stopwords to control what keywords are used for topics. To do so just add a tab "Stopwords" to your excel. You can always adjust the stopwords in the resulting excel and resubmit it. The tool always takes the first tab from the excel for its modeling.
The tool can create TopicMaps in English EN, Dutch NL, German/Deutsch DE, French FR, Spanish ES, Portugese and Greek GR. We are working on adding Russian RU.
The language is autodetected. If you need to force a different language you can do so in the Opitions (or click on the Cogs icon)
Best results for the topic map are usually achieved when named entities like personal names, cities, organisations are removed by putting them on the stopwords list. In the resulting excel file you will also find a tab with such stopwords candidates. These are the named entitites that were autodetected to the best knowledge of the model.
Realistically you can probably achieve nice results when the number of texts (rows in your excel file) are larger than 1000. This is a machine learning tool and it needs some data to learn from. Still you could also try to limit the number of topics in the resulting map again using the Opitions (or click on the Cogs icon)
Should you have problems or suggestions please contact us.