patrickbeekman / SeniorCapstone

Exploratory analysis and visualization on how to maximize the potential of your tweets.
GNU General Public License v3.0
0 stars 1 forks source link

Capstone Feedback #57

Open NaturalFractals opened 6 years ago

NaturalFractals commented 6 years ago

I would suggest breaking the technical report into paragraphs. May make it easier to read.

"This application is meant for small business social media marketers"

Background Info

Unsurprisingly enough twitter

"If users are getting more impressions on their tweets they are then likely to use twitter more which is what twitter wants so they can increase their ad revenue"

"For example the time that a tweet is sent may not be an accurate measure in my analysis because twitter will choose where to place the tweet in a users timeline affecting the time that it is seen by the users followers."

Problem scope and features

"It gives the user an easy way to aggregate all of their followers and view interactive visualizations of these general trends."

Instead of starting the last three sentences with "It does not" I would try to use different language to make the last part of the paragraph flow better.

Nice UML!

Design, Development and Testing

In a couple places twitter isn't capitalized.

Nice looking table with libraries used!

"The development of this project has been a roller coaster ride of productivity and confusion."

NaturalFractals commented 6 years ago

Development, Design, and Testing

"api"

I transitioned and got a working website with my plots in a single night, which is much quicker and less of a pain than the bulky MEAN stack.

Again, I would suggest breaking up the huge blocks of test into paragraphs so they are easier to read.

"For example testing that my application handles users who have never tweeted, I couldn’t just pick a random user who had never tweeted and hoped that they never did..."

API should be capitalized since it is an acronym.

"I mainly tested the helper methods and less so the control flow methods"

NaturalFractals commented 6 years ago

Results

"Looking at the graph we see that on the weekends there are substantially less tweets (about 16% less) which might be an indicator that there are less people on twitter then as well. "

"This is because it treats each users tweets equally, if one user has a lot of followers who can favorite/retweet their tweets and they so happen to tweet at specific hours then that will appear in the graphs and skew the analysis."

Features

Nice feature table

I had to read over this sentence a couple times, maybe it was just me...

"If I took the time to optimize my code and remove code blockers I think I could get it to run in about half the time as it takes right now."

Problems encountered

After I got to point in my project where I could download and analyze the tone of tweets I got stuck and just did some exploratory analysis for two to three weeks before finally deciding on my current analysis.

"Another problem I ran into was with all the MEAN stack stuff."