Data Scientist - Extracting Data Insights from Social Media Data - Text Categorization

We are looking for talented Data Scientists: We've successfully scraped social media data comments, sub-comments from the Niche of those who are on a specialized diet. Now need to categorize it by tags, tokenize it so we can get value out of it. We have some preliminary tags that we think that the categorization still needs some work. We need to categorize it by: - Sentiment (positive / negative) - Context (question / comment) - Problem (tags: type of problem e.g. low energy, over weight, tired, fatigue) - Disease (tags: type of disease e.g. arthritis, gout etc) - Results (tags: types of results e.g. weight loss, energy level) We're still working on the best way to categorize / tag up our data but would love to discuss with you some simple / correct ways to do this to get this project done right. Our ideal output is to understand the challenges, problems that people are having inside this Niche on a specific Carnivore Diet so we can take qualitative data and be CLEAR about what we need to focus on for our product (i.e. what problems to solve) The data is scraped from social media channels comments and sub-comments. The outputs I am looking for in this niche: 1. What are the major problems, challenges, diseases, issues people are facing 2. What are the major questions people are asking (category e.g. nutrition, food, cooking, exercise, eating times - whatever)... 3. What are the major RESULTS people are achieving, in which categories weight loss, curing inflammation, gout)... 4. Extract all records for those who have achieved GREAT results and to use as testimonials eventually .... Essentially this is Market Research . I suspect we'll have 75,000-125,000 social media comments (and sub-comments) We want to know what the major conversations are that are dominating that space so we can identify the bets product / service to build That requires us to know the problems in the marketplace, have them correctly categorized It might be useful to allow us to see the path from unstructured data > to > categorized and useful data that can be used for analysis and making business decisions. So give some thought thought through the overall project and post here your best view of how to approach this. Take it systematically and almost in sequence 1. > 2. > 3. > 4. > 5. etc. Through to tagged, labelled and meta-labelled data. You're ideal approach specific approach that will give us the resulting outputs we need