The technology behind "The Best Country for observing the Northern Lights”

The technology behind "The Best Country for observing the Northern Lights”

Denys Makogon

The principal member of technical stuff
Java Platform Group
Oracle
Date: October 24th, 2020

Transcript

1. Copyright © 2020 Oracle and/or its affiliates. The technology behind "The Best Country for observing the Northern Lights” Principal member of technical stuff
 Java Platform Group Oracle
 Date: October 24th 2020 Denys Makogon
2. Introduction and many more… Denys Makogon
3. I write software for living
4. I do landscape photography for fun, as a hobby
5. Winner Diploma awarded to Denys Makogon Winner of the 5TH annual international photography contest 35AWARDS. As a result of the primary vote photo "Pilot whale massacre. Faroe Islands." (nomination "Aerial photography", 3777 uploaded photo) entered the top 100 best photos in the 2019. Total participation has taken 117541 people from 173 countries. A total of 418 thousand photos have been submitted. Certificate № 352019436864 Powered by TCPDF (www.tcpdf.org)
6. You’re here to listen a story of how a hobby I do meets the work to do!
8. A friend of mine asked me question: Which country is the best to visit to witness the Northern Lights?
9. Pick the one you like the most!
10. The Northern Lights phenomenon is all numbers and probabilities!
11. Maybe I can do some math to answer this question?
12. Within 9 latitude belts of the Polar Circle, find a country with the highest number of night hours, between 9PM to 6AM, with the highest magnetic field disturbance (above the threshold) through the polar winter (September to April).
14. https://github.com/northern-lights-awards/awards
15. HTTP API for historical weather forecasts datasets with Solar activity and magnetic field yearly records data
16. Weather Location Season Key aspects
17. Which technology I personally found the most comfortable to work with?
18. Sequences 8 kP measures per day within “winter season” 9 weather forecasts reports per day (night hours only, 9PM to 6AM) >18.3K measures per year At least 4 countries 9 or more GPS coordinates
19. Sequences A country is a set of GPS locations Country.forEachGPSLocation()
20. Sequences A monthly statistics is a combination defined by GPS location that contains 9 multiplied by (at least) 28 weather forecasts filtered by suitable weather and magnetic field disturbance criteria
21. Sequences Pipelines Country.forEachGPSLocation() .forEachNightHour() .Monthly(September) .getWeatherForecast() .filterBykP() .filterByWeatherCriteria()
22. Sequences Pipelines Parallelism Country.parallel() GPSLocation.forEachNightHour() .Monthly(September) .getWeatherForecast() .filterBykP() .filterByWeatherCriteria() GPSLocation.forEachNightHour() .Monthly(September) .getWeatherForecast() .filterBykP() .filterByWeatherCriteria() GPSLocation.forEachNightHour() .Monthly(September) .getWeatherForecast() .filterBykP() .filterByWeatherCriteria()
23. Sequences Pipelines Parallelism IO framework
24. Sequences Pipelines Parallelism IO framework Built-in types consistency
25. And the winner technology is…
27. Streams API
28. Parallel execution Java can run multiple threads a number of parallel threads depends on a number of available CPU cores
29. S.parallelStream().operation(F) Parallel execution size_of(S) * cost_of(F) *credits to Doug Lea
30. Effectively splittable data “If it takes longer to partition data than to process it.” *credits to Doug Lea “Data structures tend to be efficiently splittable if they internally support random access, efficient search, or both.”
31. Project Loom Enhanced Java concurrency with Parallel execution virtual threads
32. Memory efficiency store as a collection | process as a stream a stream has no storage
33. Streams API dictates a code structure collection to a stream | a pipeline of processors a collector definition Comfortable API interface
34. Streams are reusable Comfortable API interface “even though the whole idea of the Stream is that it's once-off.”
35. Function<Collection<String>, Stream<?>> f = (Collection<String> c) -> c.stream() .parallel() .filter(s -> s.startsWith(“a")); Supplier<Stream<String>> f = () -> Stream.of("d2", "a2", "b1", “b3","c") .filter(s -> s.startsWith(“a"));
36. Java is great for multi-threaded data processing pipelines of both IO and CPU-bound operations. for both dependent and independent data points
37. “ *season 19-20 It’s started with a question…
38. denis_makogon java denis.makogon
39. Questions?