Data analysis of unstructured data that have non linear Key value position with Power BI

[gillgs]

I have unstructured data consist of three excel sheet

Text contents of 3 tables some how related to each other.

But there is no linear direct relationship between them e.g First row of table 1 is not related to first row of table 2 and  first row of table 3 although they sharing same numeric and letter key but may be there is relationship between First row of table 1, second row of table 2 and first row of table 3. so table 1 is related to both table 2 and table 3 via sharing common numeric key and letter key but not in orderly way.

You can consider Table 1 as questions table each row is individual question and Table 2 is Responses from one group of people and table 2 is response from other group of people. But in collected responses some people did not reply the answers orderly way . for example they Gave answer of first question then skip the answer of second, third and wrote answer of fourth one. all of data was collected on paper. so they just place bullet points in front of answers  without specify the the answer belong to what number of question.

I have to figure out some insight from data, so what kind of data analysis orapproach should I use to analysis this unstructured data with power BI. What kind of other AZURE machine learning  analysis i can use with Power BI  to figure some insights from this data.

Thanks

[v-yuta-msft]

Hi gillgs,

Which kind of data model and insight do you want to achieve? In addtion, to implement data clean or  to use machine learning algorithm, you can use R script. Also, here is azure machine learning algorithms for your reference: https://docs.microsoft.com/en-us/azure/machine-learning/studio/algorithm-choice

Regards,

Jimmy Tao

[Anonymous]

Don't know if this will work,

You can do a cross join on 3 tables with both Numeric key and letter key then you will get all possible combinations then put some conditions that will eliminate rest of records.

Thanks

[gillgs]

Hi 🙂 

But if i do cross join then may it eliminate the rows that have actual relationships and join the rows of three tables that have  no realtionship at all. i also edited my original question for more details. Please check, i think without machine learning tool from azure it is not possible to figure out the right insights 

Thanks for Reply