Twitter Sentiment Analysis using Logic Apps and Text Analytics

Question

Dear Team,

This is my first post in the Azure community. Please excuse me for any mistakes. I am new to Azure and working on one research project.

I want to analyze tweets from Twitter during the Superbowl 2021 event. I need the help with following questions.

Project Background:

1. I have approximately 50 brands (with their unique hashtags), which are participating in their brand advertising during Superbowl. I want to analyze the Twitter sentiments before, during, and after the Superbowl event and then report the analysis results to these brands.

My Current Logic App Design:

1) Created Logic App which will trigger after every 3 minues and will search for mentioned hashtags in searchtext window.
2) The tweet will then go through the Cognitive service "Azure Text Analytics" to get the sentiment score.
3) Once the sentiment score is computed, the selected fields will be stored in Azure SQL Database for further analysis
4) For analysis, the exported CSV will be used outside Azure (Analysis mostly by Python Programming later)

Issues/Requesting Suggestions:

1. Unable to capture tweets for more than 1 hashtag at a time. If I mention multiple hashtags, then the trigger doesn't capture the tweets if "ALL" of those hashtags are not present. I want to capture tweets in "OR" fashion. i.e. if any hashtag is present among a list of 50 hashtags, it should store the result.
2. Which pricing model should I select provided that I just need a maximum of 1 week of storage. Currently, I have selected vcore with 2 cores and 50GB storage (+15 GB Log storage) while creating Azure SQL Database. I did not use the DTU pricing model since I am not sure if it will be able to handle the high velocity of tweets during Superbowl. (i had selected Standard s2 with 50DTUs just to check, but the price difference is not significant.)
   Will my vcore configuration be able to handle the tweets processing? I can reduce the frequency to check tweets from 3 minutes to say 30 minutes or 1 hour. My main purpose is to store those without any CPU issues.

I would really appreciate your response!

Thank you,

Priyal