@holgerbrandl any idea why that kind thing happens?

java.lang.ClassCastException: krangl.SimpleDataFrame cannot be cast to krangl.DataFrame

Hi folks! Anaconda are doing a Data Science survey. questions about programming languages include Kotlin, if anyone wants to represent, here's the link: https://twitter.com/anacondainc/status/1102984667442147328

I usually recommend to stay away from "data science" books and instead focus on books on specific matters.

@altavir that reminds me, remember the customer waiting queue simulation you helped with? I'm going to try and make the animation continuous... if I can get JavaFX to cooperate.

I think I'm going to use that for my presentations this year. I'll make sure to credit you.

My latest article about questioning the supremacy of Python in Machine Learning followed by a state of the ecosystem in Kotlin and an example in Spark. https://www.thomaslegrand.tech/kotlin/2019/04/07/machine-learning-kotlin.html Slack Conversation

If anyone needs an automatic differentiation engine in Kotin you can use this gist of mine (I don’t have time to turn it into an actual library). It is not the most optimal implementation in terms of performance, but it is quite easy-to-use in terms of API and has very concise and simple implementation + Kotlin DSL FTW. https://gist.github.com/elizarov/1ad3a8583e88cb6ea7a0ad09bb591d3d I’ts backwards-mode (back-propagation), so you get derivative of one output value by all input values in one pass. That is the thing you’d need to do gradient descent. Use it like this:

val x = D(2) // define variable(s) and their values
val y = grad { sqr(x) + 5 * x + 3 } // write formulae in grad context
assertEquals(17.0, y.x) // the value of result (y)
assertEquals(9.0, x.d)  // dy/dx

Hi, I'd need a weighted random operation. I saw Thomas'

Kotlin Statistics

includes it (and it's the one I'll probably use) but I am a bit confused about the libraries I read about here. Would

KMath

include this kind of operations in the future? Should I look into some other lib? Thanks!

If you all can help me with a poll. https://twitter.com/thomasnield9727/status/1134493388236046336?s=09

There is a term describing the fallacy of equating a video game or board game AI with having real world applications and AGI capabities. For example, a chess algorithm that wins chess does not mean it can master strategy in corporate or military environments. Does anybody know what this term is? Reason I ask is because I'm considering writing about the fetish of video games in AI research. https://deepmind.com/blog/alphastar-mastering-real-time-strategy-game-starcraft-ii/

is there anyone who has generalized serialization for hazelcast (or geode, or ignite? which i have yet to try out)

Is there a place I can start looking at?

I am sick and tired of python, and there's nothing comfortable about it, but without CUDA or seriously tuned opencl backend I'm not going to waste my time training models at 1/50th the output

Someone asked about kotlin for data sceience? Look here: https://www.jetbrains.com/careers/jobs/senior-java-kotlin-developer-kotlin-for-data-science-288/.

Data-science friends, what do you think of this? https://kotlinfrompython.com/2017/10/29/when-will-your-project-grow-up-and-require-typesafe-code/

people in this discussion seem to be like perl coders blissfully satisfied with thier methods and progress in accomplishing their goals, generally the question is "why do we need any less restrictive syntax"

By the way, I plan to move pandas-like table API from my old project. I won't implement it inside kmath, but rather inside dataforge - more sophisticated piece made for more high-level data interaction (https://github.com/mipt-npm/dataforge-core/issues/19). I plan to proceed the same way as in kmath, meaning focusing on API, providing only basic implementation and then linking with existing libraries like tablesaw, krangl and maybe even pandas itself, when I finally get to native. It would be really good to discuss. what do we need from table API.

kotlin has a better "score" than java, to be sure, for most situations. java is low entropy. The reason why Java is viable at all is because in early 90's, c++ was dominant, and there were a ton of limitations to the available late binding options; in c++ at the time, progress was generally a code-generation tool or or a precompiler of c++ extensions which emitted c++ virtuals. I don't think that the jvm evolution should stop with the current decisions of kotlin, personally. data science is heavily dependant on information thoery and to ignore the signposts where a given solution depends on approaching Shannon's limit is to reverse the efficiency of the solution in question. intelligence is based on comprehension. comprehension is the orthogoanl arrangement of information to align meaning with content. code is ultimately an expression of data.

@Amir Gur @altavir this is not directly the application of kotlin to big data -- however, my involvement with this has been to encourage implementation towards large first-order-predicate logic systems over the years. I'm casually refactoring this to Kotlin with the main author's blessing. https://kotlinlang.slack.com/archives/C2YH04E3S/p1571505768009700

@altavir i have bookmarked kmath, fwiw. I still do wish the recipes were nearer to the front tho.

i have a common usecase: really common. I have a table from csv: Date, Customer, ItemId, qty_sold, qty_returned what is the kmath recipe to pivot distinct ItemId to new columns Date, Customer, qty_sold_ItemId1 ,qty_returned_ItemId1, qty_sold_ItemIdN, qty_returned_ItemIdN,...

@jimn Wait is that a way to read python dataframes in kotlin?

im not convinced you are working with large databases frequently from your comments. there is no simple option to harvest diverse legacy production database into some shiny new faster BI platform, that's consulting golden-handcuffs snake-oil. how would you move a 300 gigabyte relational database to bigquery or amazon for instance?

never tried the java bindings

did you try https://wiki-bsse.ethz.ch/pages/viewpage.action?pageId=26609113

I see…

Hi all 😃 I was looking at the kotlin-statistics project and the histogram implementation. I have this problem that I hope you have some input on. The idea is that in a calendar the business can set price points and the price will change night and weekend and holiday's even down to a minute precision. When the customer has done it's trip we have the interval in time and my idea was that I read up the price points (have not figured out how to store it yet) into a histogram and then lay over the interval over the histogram and we then get a total price and a price per bucket(pricepoint) Any thoughts on this? And is the kotlin-statistics good for this problem? Thanks 🙂

Enjoyed @altavir’s talk, especially the use of generics and scoping for type safe numerics: