Goethe University Frankfurt

lightbulb-2692247_640

Data Challenges 2018

(Summer Semester 2018)

B-WB, M-WB, M-PoE

Lecturers: Prof. Dott. Ing. Roberto V. ZicariDr. Karsten TolleKim HeeTodor Ivanov and Naveed Mushtaq

Mentors:  

Name Topics
Daniel Amthor technology, business models, smart cities
Adam Azani startups, innovation, business models
Prof. Nils Bertschinger machine learning
Björn Braun new technologies, design thing / UCD, market research,  business model design
Jonas De Paolis winner of ING-DiBa Data Challenge 2017, technology, business models
Sead Izberovic data security, data management, technologies
Prof. Dr. Udo Kebschull data security, networks
Alex Klein technology, web applications, software engineering, startups, innovation
Klara Kletzka social, societal innovation
Patrick Klose winner of the DB Data Challenge 2017; Artificial Intelligence (esp. Reinforcement Learning) and Software Development
Hevin Özmen urban transportation
Nicolas Pfeuffer winner of DB Data Challenge 2017; Artificial Intelligence (esp. Conversational Agents) and Software Development
Dr. Manfred Spindler business angels Frankfurt, business models, innovation
Prof. Roser Valenti innovation, research
Rut Waldenfels transport, business models

Course start: Thursday 19. April. 2018 (kickoff)

Time and Location:

Thursday  14:00 – 16:00   Hörsaaltrakt Bockenheim – H III,  Gräfstraße 50-54, Frankfurt. (Campus Bockenheim), Goethe University Frankfurt. See map here: http://www.bigdata.uni-frankfurt.de/about/

Friday       10:00 – 12:00   Hörsaaltrakt Bockenheim – H II,  Gräfstraße 50-54, Frankfurt. (Campus Bockenheim), Goethe University Frankfurt. See map here: http://www.bigdata.uni-frankfurt.de/about/

Languages: The languages of the lecture are English and German.

Credit Points: Students can receive 6 CPs. Link in QIS/LFS

Course Description: Students will take part to two Data Challenges. One offered by Deutsche Bahn AG and one offered by Procter & Gamble (P&G).

Eligibility: Bachelor Students, Master Students, and PhD students across multiple disciplines are encouraged to attend the kickoff and to sign up for one Data Challenge.

Students in Computer Science, Data Science, Information Systems, Business Computer Science, Mathematics, Economics, Marketing, Psychology, and other disciplines will form teams of two to explore the questions posed.  Team members are required to attend the kick-off lecture to sign-up for this project.


Important Note: This project is in two phases, with Phase One and Phase Two taking place Summer Semester 2018.  

Among the teams that successfully complete both phases of the project, winners will be awarded a prize.


lightbulb-2692247_640 Deutsche Bahn and Frankfurt Big Data Lab Data Challenges 2018:

“Smart Cities, Smart Life”

 

The design of our future urban life will depend fundamentally on the constitution of our future mobility in ever faster growing or ever faster evacuating urban regions.

The Deutsche Bahn AG, as one of the few comprehensive mobility supplier, wants to shape the actual development towards networked, smart and sustainable mobility.

Against this background the development and the construction of digital and networked possibilities to offer comfortable, payable and simultaneously environmentally friendly solutions of urban mobility and logistics will gain great importance.

This means in our opinion the following:

  1. How will be organized smart and integrated mobility on different modes of transport in the future
  2. How will be shaped smart logistics and how could we offer new infrastructural services
  3. How will become reality a new smart mobility by the use of nodes and transfer points and the remodeling of third places
  4. Which concepts for a „mobility on demand” or certain possibilities of autonomous driving actually could be demonstrated prototypically

On the basis of the open Data pool of Deutsche Bahn AG, students of the Frankfurt BIG DATA LAB should create ideas and patterns of possible solutions that could accomplish to be regarded as possible indications for a smart mobility of the future.

For first insights on actual solutions look at the following reference implementations from Deutsche Bahn AG:

  1. Bahnhofsbox
  2. Cargo Bike
  3. Clever shuttle
  4. ioKi

DB Challenge Prizes:

  1. Weekend trip to Berlin with visit of our DB MindBox + Gold Trophy
  2. Voucher of ICE (class 1) railway trip to a main city destination in Germany + Silver Trophy
  3. Value Voucher & All-in-One Charger + Bronze Trophy

 

lightbulb-2692247_640    Procter & Gamble (P&G) and Frankfurt Big Data Lab  Data Challenges 2018:

“Smart Logistics, Smart Supply Chain”

 

 

At P&G, everything we do starts by winning with consumers and shoppers.

Our aspiration is to serve the world’s consumers better than our best competitors, in every category and every country where we choose to compete — creating superior shareholder value in the process. P&G is focused on four key areas of transformation to deliver balanced growth and leadership value creation: streamlining and strengthening our product portfolio, improving productivity and our cost structure, building the foundation for stronger top-line growth, and strengthening our organization and culture. We are organizing our portfolio around 10 product categories and about 65 brands — approximately half of which have sales of more than $500 million each year.

In a dynamic manufacturing and retail environment, data and analytics offer unique opportunities to better understand and serve consumers. Shoppers expect personalized experiences – just the right information and inspiration, tailored to their needs, and the products that they want, easily available offline & online. Are you up to the challenge to transform our go-to-market and serve consumers and shoppers in the Frankfurt area best?

Bring your expertise in data and technology, your curiosity to discover the world of fast moving consumer goods, and your innovative ideas to delight consumers. The challenges we want to address are:

  • How can we leverage historical data and self-improving algorithms to better predict future product demand?

The focus will be on one or a set of German customers. You will have access to historical ordering and shipment data, supply chain details (distribution centers, retailer outlets) as well as consumer demand (product offtake). Other factors (such as weather forecast) might influence the demand for certain products. The algorithm should predict the future order volumes and patterns, and ideally be able to learn from new data being introduced. The outcome will be an improved supply chain, transport efficiency and increased shopper satisfaction via reduced out of stocks.

  • How should the retail landscape transform itself to better serve consumer demand in Frankfurt?

Frankfurt is a dynamic and diverse city. We know the current retail landscape (outlets location and types of stores) as well as demographic attributes of the different neighborhoods. Are shopper needs currently being met? How should offline and online shopping offerings evolve to meet the ever-changing demands?

P&G Challenge Prizes:

  1. 2-days trip to Geneva incl. visit of P&G headquarters + Gold Trophy
  2. 2-days trip to Cologne incl. tickets for DMEXCO and visit of a P&G plant + Silver Trophy
  3. P&G Product Prize (e.g. OralB Genius Toothbrush Set) + Bronze Trophy

Project Description:

The project consists of two phases: Phase I and Phase II will be held during the Spring Semester 2018. The proposed timeline and details of these stages are:

Phase 1:


-Teams will be asked to address one of the Data Challenges offered.

Specifics will be addressed at introductory lectures. Teams will then work independently to create a proposal of a novel idea that satisfies the data challenge chosen.

-Deliverable: A mid-term presentation of the project idea, where it is required that:

  1. teams clearly state objectives,
  2. general description of the way they intend to implement the idea using the data available for the challenge chosen.

Phase 2:

Teams that submitted a successful presentation at Phase I will be then asked to implement the idea and present it at the end of Phase II.  (Exact dates and detailed agenda to be reviewed at the kickoff)


Course Schedule (preliminary)

Date Topic Materials
19.04.2018 – 14:00-16:00 Kick Off Meeting Data Challenges: Deutsche Bahn and Procter & Gamble (P&G)
26.04.2018 – 14:00-16:00 Data Challenge presentation by Deutsche Bahn
27.04.2018 – 10:00-12:00  Data Challenge presentation by Procter & Gamble (P&G)
03.05.2018 – 14:00-16:00 Jürgen Kohnen (P&G)  ‘empowerment training’ + Q & A
04.05.2018 – 10:00-12:00 How to do: business understanding, requirement analysis + How to handle data? – Data Tools + Q & A
11.05.2018 – 10:00-12:00  Meeting the Mentors +  Q & A
submission deadline for P&G Challenges: 14.05.2018 – 23:55 Phase I – by e-mail to: dc@dbis.cs.uni-frankfurt.de
17.05.2018 – 14:00-16:00 No Lecture
18.05.2018 – 10:00-12:00  Student presentations – Procter & Gamble (P&G)
submission deadline for Deutsche Bahn Challenges: 21.05.2018  – 23:55 Phase I – by e-mail to: dc@dbis.cs.uni-frankfurt.de
24.05.2018 – 14:00-16:00   Student presentations – Deutsche Bahn
25.05.2018 – 10:00-12:00
01.06.2018 – 10:00-12:00
07.06.2018 – 14:00-16:00
08.06.2018 – 10:00-12:00 P&G Challenge 2 – meeting with Roland (at DBIS)
14.06.2018 – 14:00-16:00 P&G Challenge 1 – meeting with Torben (at DBIS)
15.06.2018 – 10:00-12:00 Meet the Mentors + all student teams
21.06.2018 – 14:00-16:00
22.06.2018 – 10:00-12:00 Milestone check for all teams + DBIS
28.06.2018 – 14:00-16:00
29.06.2018 – 10:00-12:00 Meet the Mentors + all student teams
05.07.2018 – 14:00-16:00 Presentation on Smart Mobility
06.07.2018 – 10:00-12:00
12.07.2018 – 14:00-16:00 Final presentations Deutsche Bahn Each team will have 15 minutes (+ Q&A) for final presentation.
13.07.2018 – 10:00-12:00 Final presentations Procter & Gamble (P&G) +  Award Ceremony at Procter & Gamble (P&G)  Each team will have 15 minutes (+ Q&A) for final presentation.
Award Ceremony at Deutsche Bahn

 


Resources

UC Berkeley DATA-resources – many course materials on Python, NumPy, Pandas, SciKitLearn, MatPlotLib, TensorFlow, Machine Learning and more

Mobility

Ethics and Data

Legal Implications of Data

Data Privacy

Elevator Pitch

Elevator Pitch- 5 minutes Presentation

Machine Learning

  • Machine Learning Course at Stanford by Andrew Ng, Chief Scientist of Baidu; Chairman and Co-Founder of Coursera; Stanford CS faculty.
  • Non technical 5-part series on introductory machine learning by Alex Castrounis, Product Leader and Technologist.
    • Part 1 – definition of machine learning and most widely used machine learning algorithms.
    • Part 2 – model performance, data selectionpre-processing, splittingfeature selection and feature engineering.
    • Part 3 – model variancebias, overfitting, model complexitydimensionality reduction, model evaluationperformance, tuningvalidationensemble learning, and resampling methods.
    • Part 4 - model performance and error analysis 
    • Part 5 – unsupervised learning, predictive analyticsartificial intelligencestatistical learning, and data mining.
  • Downloadable CRC Press Free Book on “Explorations in Artificial Intelligence and Machine Learning” (LINK to CRC Web site- registration required) with 7 chapters:
    • An Introduction to Machine Learning
    • The Bayesian Approach to Machine Learning
    • A Revealing Introduction to Hidden Markov Models
    • Introduction to Reinforcement Learning
    • Deep Learning for Feature Representation
    • Neural Networks and Deep Learning
    • AI-Completeness: The Problem Domain of Super-intelligent Machines

Open Source Tool

  • Apache Hadoop is a project developing open-source software for reliable, scalable, distributed computing.
  • Apache Spark is a fast and general engine for large-scale data processing.
  • Apache Flink is an open-source platform for distributed stream and batch data processing.

Advanced AI Tools

  • TensorFlow  is an open source software library for numerical computation using data flow graphs.
  • The Microsoft Cognitive Toolkit: A free, easy-to-use, open-source, commercial-grade toolkit that trains deep learning algorithms to learn like the human brain.

Making App

Chat Bot 

Enterpreneurship

(C) Big Data Laboratory. Design By Tea Sets