AlgoIL #10 [In Rehovot!!] - Locating Buses and Bus Stops

AlgoIL is visiting Rehovot!! [10 min. walk from the train station!]
Join us for a special event in Rehovot where we will listen to two fascinating talks about unexpected complications in the world of smart transportation :)

Agenda:
17:30 Gathering and snacks
18:00 Opening words + Hello from our Hosts
18:05 Itay Ovits - Vehicle Location Estimation, and Making It Work in the Real World
18:50 Short break
19:00 Rotem Duani - Detecting Bus Stop Locations? Not as Easy as You Would Think
19:45 Time for more snacks and chats

Thanks Powerball and Hip for hosting this event!
[Notice the venue is only 10 min. walk from the train station!]

Both talks will be held in Hebrew.
Tha talks will be filmed and uploaded to our YouTube channel: http://bit.ly/AlgoILYoutube
_____________________________________________________________________________________
_____________________________________________________________________________________

Itay Ovits - Vehicle Location Estimation, and Making It Work in the Real World

Everyone needs to get to work. But for many people, especially those who live in the suburbs, commuting is the worst part of the day.
As part of Hip's smart-transportation product, which aims to improve commuting, we had to determine the exact location of each of our vehicles in real-time, especially with respect to the pickup and dropoff locations.
Now, I know what you're thinking: yeah, been there, done that. It's a solved problem!
However, as is often the case, the constraints we had and the tradeoffs we chose, made this a far more complex problem than we initially thought.

In this presentation I'll talk about the problem, the issues we've encountered and how we solved them; about practical considerations in choosing the algorithmic / data-science-y solutions, especially with respect to engineering work and future maintenance; and about lean approaches to modelling (hint: your data is the model).

About the speaker:

After almost 10 years in various R&D positions at 8200, I've moved to the start-up world - first in research positions, and later in managerial roles and as a founder. I'm the CTO and co-founder of Hip, where I apply both my algorithmic skills and management experience to complex real-world problems, aiming to improve everyone's commute.

_____________________________________________________________________________________
Rotem Duani - Detecting Bus Stop Locations? Not as Easy as You Would Think

Part of the fun of developing an algorithm is starting with a classic algorithm and then figuring out which adaptations and faults it has in order to modify them so they will fit and solve our problem.

In the presentation, I will demonstrate one challenge we solved during the early days of Moovit. When we started mapping bus stop locations, we realized that the data we received from external sources were wrong. We developed an algorithm that will automatically locate the stops correctly. I will give a quick overview of our data, and present the original solution and its evolutions during the research, resulting in an efficient algorithm that now operates automatically around the world in 3,000 cities.

About the speaker:

Rotem Duani is a data scientist and a part of the Data Science team at Moovit, who researches and develops many algorithms at the core of Moovit’s B2C and B2B products. Rotem holds an M.Sc. in Intelligent Systems and Statistics and a BSc.

_____________________________________________________________________________________