James
jamesgibbins.com
James
@jamesgibbins.com
항상향상
🇬🇧📈🏍️🎧⚡👊🏻🐕🐄
3/5: There’s a pretty clear correlation between being religious and hating homosexuals. I thought religion is about love?

datacanvas.substack.com/p/religion-a...

#Religion #Tolerance #Culture #HumanRights
November 4, 2025 at 9:58 AM
5/5:
September 24, 2025 at 8:42 AM
5/5: In 5 English towns, over 1/3 of people do not speak English well. 1 in 70 foreign-born residents in England and Wales speak no English.

www.dailymail.co.uk/news/article...

#Language #Culture #Immigration #Multiculturalism #Assimilation
May 28, 2025 at 9:53 AM
3/5: Distort the map using various projections by centring different countries. Australia Mercator is a fun one - Russia becomes terrifying.

engaging-data.com/country-cent...

#Maps #Projections #Travel #Perspective
May 28, 2025 at 9:47 AM
5 #shareworthy things

1/5: Zooooom all the way in. It’s almost like a natural landscape.

www.micro-pano.com/pearl/index....

#Art
May 28, 2025 at 9:44 AM
5 #shareworthy things:

1/5: Your guess for how long something will take is the median. However, it’s lognormal, so the mean is 1.6x that, and the 95th percentile is 5x. Don’t tell someone it will take take 4 hours (half a day), say 20 (2.5 days)

thesearesystems.substack.com/p/task-estim...

#Stats
May 20, 2025 at 10:12 AM
#Netflix beats #YouTube when searching "you" on #Google. Pretty impressive from the @netflix.com team.

#SEO
February 7, 2025 at 7:37 PM
4/5: The UK is the same latitude as Canada. Without the Atlantic Meridional Overturning Circulation (Amoc), the temperatures may be more similar too. Not guaranteed, but a possibility.

@financialtimes.com #climate #climateChange #brrr #England #UK #AMOC #gulfStream

www.ft.com/content/7711...
January 22, 2025 at 10:59 AM
Time for something different - a dashboard! While Python is great for transforming and analysing data, it’s not the most end-user-friendly. So, I made an interactive Power BI dashboard.

www.jamesgibbins.com/dvla-data-da...

#DVLA #DataAnalysis #BusinessIntelligence #BI #Dashboard #PowerBI
January 20, 2025 at 11:19 AM
I analysed #DVLA vehicle registration data to answer:

1. What are the most common cars in the UK?
2. Which old vehicles are having a comeback? (Might be an investment opportunity)

Some results were genuinely surprising.

www.jamesgibbins.com/analysing-dv...
#DataAnalysis #Python #Cars #ClassicCars
January 17, 2025 at 9:33 AM
Another little project - send a word or phrase to an API, and it will give you a sentiment score between -1 (negative) and +1 (positive).

www.jamesgibbins.com/basic-apis-s...

#DataScience #DataAnalytics #NaturalLanguageProcessing #NLP #SentimentAnalysis #API #Python #TextBlob #Flask #Docker #AWS
January 8, 2025 at 8:51 AM
One of the podcasts I listen to most, The @jordanharbinger.bsky.social Show, has Feedback Friday. I felt it has become too negative. Can I support this hypothesis with data?

jamesgibbins.com/podcast-sent...

#DataAnalysis #NLP #SentimentAnalysis #TopicAnalysis #ThemeAnalysis #LDA #Python #Podcast
January 7, 2025 at 8:05 AM
Like many people, I was a bit disappointed with my Spotify Wrapped this year. So, I thought I’d create my own. You can generate the insights with your data using my code too, if you want.

jamesgibbins.com/spotify-wrap...

#DataAnalytics #Python #Spotify #SpotifyWrapped
January 6, 2025 at 7:40 AM
Natural language processing (NLP) involves getting a machine to understand human language. I made a model to predict whether a tweet referred to a genuine disaster, potentially helping first responders.

jamesgibbins.com/disaster-twe...

#DataScience #DataAnalysis #NLP #Python #Twitter #SocialMedia
January 2, 2025 at 2:52 PM
Now for something a bit different! A SQL-based murder mystery. Simple, but fun.

jamesgibbins.com/sql-murder-m...

#SQL #DataScience #DataAnalysis #Northwestern #KnightLab
December 18, 2024 at 1:42 PM
Part 4 of 4. Revenue and sales per brand and per customer group, then elasticity of price and quantity. Raise or lower prices? Do promotions help?

#DataScience #DataAnalysis #Python #MachineLearning
#CustomerAnalysis #MarketAnalysis #PricingStrategy

www.jamesgibbins.com/customer-ana...
December 11, 2024 at 9:31 AM
Part 3! Linking purchase data to our clustered customers for more insights. Top customers? Revenue? Visits-purchases correlation? Trends over time? This, and more!

www.jamesgibbins.com/customer-ana...

#DataScience #DataAnalysis #MachineLearning #Python #CustomerAnalysis #PurchaseAnalytics
December 5, 2024 at 2:01 PM
Continuing my customer analysis project - classification. K-nearest neighbour, random forest, XGBoost, logistic regression, and more.

www.jamesgibbins.com/customer-ana...

#DataScience #DataAnalysis #MachineLearning #Python
#CustomerAnalysis #MarketAnalysis #Classification
December 2, 2024 at 11:28 AM
So it's mostly chosen to make the graph look good? It just seems a bit weird to me, a bit like p-hacking, choosing a period for that reason. I always assumed the plot was more absolute, like this one from Wikipedia.
November 27, 2024 at 3:28 PM