Clojure, a Sweetspot for Analytics

on YouTube ↗37m

" summary

Describes a Clojure backend that handles millions of messages per second and shortens time-to-production, using Anglican probabilistic programming to model latencies and priors (coin-flip examples with 50,000 trials, 28k tails) and a statistica + core.matrix stack for Naive Bayes and linear regression (using Apache Commons Math for gradient descent). Shows a Cassandra latency simulation where 500,000 requests drive latency to seconds and identifies ~30 nodes as the practical scaling point.

AI-written summary · about

" chapters (8)

1. 0:00Start
2. 3:18Why Clojure is so awesome?
3. 11:28Getting something out of nothing
4. 11:56understand the shape of data in a pace that suits your business
5. 23:10Naïve Bayes Algorithm
6. 28:34Implementing Linear Regression with Gradient Descent
7. 32:30Bunch of JVM libraries available clojure.matrix is great clojure.match greatly helps with algos Clojure fins are easy to test With immutable DSS nothing goes wrong
8. 34:31Conclusions

" related talks

• 33mClojure Data ScienceEdmund Jackson · Clojure/conj 2012
• 33mClojure Dispatches From a National Research LabJennifer Lewis · Clojure/west 2017
• 18mClojure Versus the BotnetsMarshall Bockrath Vandegrift · Clojure/conj 2015
• 37mClojure Remixed: Pyroclast Data ProcessingMichael Drogalis · Clojure/west 2017
• 36mMoving Faster: Why Intent Media Chose CascalogKurt Schrader · GOTO Chicago 2014
• 38mClojure for Business Teams - Decomplecting Data AnalysisRam Krishnan · Clojure/conj 2015
• 46mCascalog: Making Data Processing Fun AgainNathan Marz · Clojure/conj 2011
• 37mCascalog 2- Streaming MapReduce in ClojureSam Ritchie · Clojure/conj 2013