Bruce Meng (@BManalyst) header
Bruce Meng (@BManalyst) profile picture

R user! Likes making a computer do calculations, sometimes unnecessarily. Also likes cycling. Go #rstats.

Toronto
Joined August 2010
It took 5 years to build the number of jobs lost in only 4 months from #COVID19 in #Canada. Data from StatsCan. Visualization made with #RStats.
Last one for now on algos reading books: Trained a machine learning algo on ~300 books from the #ForgottenRealms universe. Algo extracted characters/locations and their relationships. @r_a_salvatore characters dominate! Made using #rstats.
4
3
This is the second in my series of algos reading books :) Here's a character extract and relationship mapping from the books in the #TheExpanse universe (not including Tiamat's Wrath) by an AI algo. Made with #rstats
2
1
Just wanted to share... I trained an AI algo to read the #GOT books; the AI algo thinks there are four main story arcs (shown as four separate colour groups). More details on this reddit post! reddit.com/r/dataisbeautiful… … Tools: #rstats, #keras, igraph/ggraph
1
3
4
Went exploring teacher misconduct text after learning that a former high school classmate was found guilty of sexual misconduct with a high school student. Made the following with #rstats. If you wish to explore the data yourself: brucemeng.ca/post/a-look-at-…
1
2
3
I added an update to autoSmry: 1. Direct url input 2. Additional languages: Fr, De, Es autoSmry is a #rstats tool I created that tries to find the most important sentences in an article. autoSmry link: brucemeng.ca/project/autosmr…
1
2
Auto neural network model vs. Manual Keras model - which is better at time-series prediction? (animation powered by @plotlygraphs). See the results here: brucemeng.ca/post/auto-neura… #rstats #DataScience
1
5
11
Just realized that Keras has a layer_cudnn_lstm which can be used in place of the regular layer_lstm... it resulted in a 85-90% reduction in training time for my application 😲 #rstats
1
2
I created a #rstats app to pick the most important sentences from an article of text in order to produce an automatic summary. It tries to do this by finding similarities between the words. It also has a neat physics-based visualization! Link to autoSmry: brucemeng.ca/project/autosmr…
1
2
1
8
Do you sometimes have trouble understanding other humans? Perhaps human-human relations droid SN2-D2 can help! 😀 brucemeng.ca/project/sn2-d2/ #rstats #rshiny
1
2
1
1
Now compare that to when size = 2...
1
Ah... I see what you are saying. Yes, it seems the line size is not a 1:1 match with the original ggplot output, but you can tweak the size of the line with the "size" argument (e.g. +geom_line(size = 0.2)). I used the size = 0.2 in the gif below:
2
1
gganimate vs. plotly... See how the two #rstats animation packages compare! brucemeng.ca/post/animations…
The #NVIDIA Titan V is... unbelievable! I've been training all my #rstats models with my CPU this whole time and... well... see the difference for yourself! I also compared the mining hashrate on #garlicoin... holy %$#! Full: brucemeng.ca/post/titan-v-an…
2
1
3
Replying to @Rbloggers
I get a profound sense of pride and accomplishment when I can get my cpu to use all of its cores... 😂 #rstats caret rocks
Fantastic birthday meal guys! Thanks for coming out!
Running a parallel model building command with caret/doSNOW just makes me smile... 😀 #rstats
1
With the holidays being so near... I did some light analysis on consumer spending in Canada. Looks like most people wait until 3 - 4 days before Xmas before doing their shopping! Full post here: brucemeng.ca/post/consumer-s… #rstats
2