The Yasna platform can process external datasets of text responses structured in a question database format. Access to this functionality is available with all plans including Free plan.
You can analyse 5000 data cases a month for free. For any feedback, feature suggestions and bug reports as well as for requesting enterprise plan please contact us at [email protected]
Have a look at a quick interactive demo to see the main features of the text processing tool
Have a look at a quick interactive demo to see the main features of the text processing tool
Data format
Before downloading the source data and processing it is necessary to format the data as follows:
- The data file should be in xlsx or csv format, using the delimiter - ‘;’
- Each column should contain response data for 1 question, and the wording of the question itself should be in the column header (= first row of the table).
- Optionally, you can add the interview ID (key) as the first column. In that case, rows with the same keys will be read as one interview. If you don’t use keys (option ”The data has IDs in the first column)” is off), the platform will assign them automatically. In this case, the data from different columns in a row will be read as belonging to one interview.
- You can also optionally add columns with variables related to the respondents in the database, so that you can make subgroups for analysis later.
Example of a table with raw data ready for processing
| ID |
What associations do you have with this name? |
What do you think this title is for? |
Title (cell) |
Gender |
| 1000001 |
swan lake ballet |
cosmetics brand |
White swan |
Female |
| 1000002 |
song |
I don't know |
Black raven |
Male |
| 1000003 |
Russian folk songs |
may be domestic clothing brand |
Black Raven |
Female |
| 1000004 |
Chanson, park, shashlik |
for vodka |
White swan |
Male |
| … |
|
|
|
|
The file can have any number of lines, questions, and variables.
Limitations of the report functionality
- We support only one level of subgroups. For example, if you need to create subgroups Gender X Age, but in the data table they are different columns, the platform will only allow you to do it by manually filtering interviews by ID list. Therefore, for optimization, it is better to mark in advance in the file, which interview belongs to which subgroup.
- On the platform it is possible to make separate reports in Storytelling format (text summary) for any number of subgroups. However, the Scale mode only allows you to create one codifier per selected array. Saving multiple codifiers for the same question is not possible, the new codifier will erase the data of the old one. If you know in advance that you will need different codifiers for cells when coding (e.g. by concept), it is better to split the data for them in advance - either into separate files (and process them as different projects on the platform), or into columns with different headings for questions by concept - see the example below:
| ID |
What associations do you have with this name? (White swan) |
What associations do you have with this name? (Black raven) |
| 1000001 |
swan lake ballet |
|
| 1000002 |
|
song |
| 1000003 |
|
Russian folk songs |
| 1000004 |
Chanson, park, shashlik |
|
| … |
|
|
- The coded response database is uploaded in R format only, if necessary, we can merge it with the numerical analysis data and convert it into any other format convenient for the client on request. Only the percentages displayed in the table can be copied directly from the platform - see below.
Working with data step by step