PyData TLV meetup #11

This will be our #11 amazing PyData event in Israel! As usual, it will include great lectures by industry experts, mingling and sharing :) All lectures will be held in English.

Many thanks to Via - On Demand Transit for sponsoring this event!

Schedule

• 18:00 - 18:30 - Gathering, snacks, mingling

• 18:30 - 18:40 - Opening words

• 18:40 - 20:20 - Three Lectures

The talks:

Speaker: Dalya Gartzman

Title: Offline Solutions to Online Problems

Abstract: We live in an online world. As consumers we expect our demands to be answered on the spot, and as creators we design our products to answer consumers' immediate requests. In Via (https://ridewithvia.com/), for example, the minute you ask for a ride, you will be assigned a van - a live imitation of a greedy algorithm. But what if the algorithm had the input that in a minute from now I will also ask for a ride with similar origin and destination? then it will be able to optimize the assignment - it will keep you waiting for just a minute, so that we can ride together, even more cheaply and economically than ever!

To address this issue, we will talk about a project I lead in Via, where the goal was to imagine we knew all requests in advance, and make the best assignment to vans based on this "omniscience". I will use this project to present a python package I used (networkx) and discovered some cool stuff about, and will share with you the exciting data-exploration of the results.

About Dalya: A mathematician at heart, an algorithmatician in practice. After receiving my MSc. in math from TAU I ventured into data science and algorithms domains, glad to discover that the real world is at least as exciting as the abstract. Currently algorithms researcher at Via - On Demand Transit, applying both my theoretical and practical passions.

Speaker: Maya Dotan

Title: A Data Scientist's Introduction to Bitcoin and the Blockchain

Abstract: Bitcoin and the Blockchain have become increasingly used buzzwords in the past few years. The value of bitcoin has skyrocketed and is currently set at 1BTC~12000$(!!!). We hear of a new completely anonymous coin that can't be forged, can't be controlled (or printed) by any government and is completely distributed. While Bitcoin does introduce a truly revolutionary concept, it is not without flaws. In this talk we will discuss some of the technical specifications of Bitcoin, as well as some limitation and vulnerabilities of Bitcoin and the blockchain. We will learn how anyone can easily explore the blockchain using a simple python package (in fact there are many such packages). We will also attempt to see how data-science (and graph theory) techniques can be used to further explore the blockchain to try and gain some insight to the nature of the usage of Bitcoin today.

About Maya: Maya completed both her BSc. in Mathematics and her MSc. in Computer Science at the Hebrew university in Jerusalem. Her Masters thesis focused on High-dimensional combinatorics, graph theory and Markov processes. In the past she worked for several tech companies, focusing on performance enhancement of large scale systems, and the efficient integration between code and large scale databases. Today Maya is a PhD candidate at the Hebrew University, her focus is parallel and distributed computing.

Speaker: Noah Dolev

Title: Which is better? Telling any two things apart

Abstract: I've been reading ML and ANN papers for a while now and it seems to me like an entire discipline based on statistical inference has decided to ignore the fact that the question of which model is better is a statistical one. Is a test set precision of 0.834 better than 0.821 in terms of its ability to characterize the model's performance in the real world? Without an estimate of the variance in precision - how can you know? I will explore the possibility of estimating this variance so we can actually pick the approach which is likely to work better.

About Noah: Name's Noah Dolev minus the ark and animals. I am a computational neuroscientist by training. During the day, I work at Wix as a mild-mannered data scientist and statistician but at night I am RunAway Man! I use my dorky powers to escape more virile males by running away quickly!