Our mission is to manage a huge amount of communication and document data in large scale industry projects by providing web based project management systems. The increasing amount of communication creates the desire for a GPS helping us and our customers to navigate through the communication stream. Our R&D projects are focusing on topics like clustering, event detection, and network analysis (Who knows who, domain experts).
Traveling the wild side of NLP, Data Science, and Analytics, we stumbled across amazing Python tools supporting us in our goal to navigate the project communication and therefor supporting our clients in Project & Risk Management avoiding wrong turns. We would like to share some of our approaches to answer our research topics and challenges:
One of the challenges, amongst others, is to utilize and adapt up to date clustering algorithms for social stream data and to expose them as reentrant services. Another one is to tailor them for the current application domain, improving clustering precision by parametrization and other means. Furthermore the integration of a Python based analytics system into an existing JAVA based application environment and eco system is required.
In addition, we would also like to share some of our “traffic jams” experienced during our travel starting as traditional Java/SQL focusing company that integrated Python into its development portfolio.
2. 14.07.2017
SOBIS | Our software and your power for successful projects.
2
From an old-school data managing company …
… to data analytics with Python
3. 14.07.2017
SOBIS | Our software and your power for successful projects.
3
» Stephan Gramlich
• Head of Products & Services
• Lecturer (IT PM) at Cooperative State
University Mannheim, Germany
• Social network perspective to IT
service management
• linkedin.com/in/stephan-gramlich-
3ab28327/
» Henrik Hain
• Software Developer & Analytics Evangelist
• M.Sc. Student @Karlsruhe Institute for
Technology
• Machine Learning, Neural Nets, Cognitive
Systems
• Research on event detection in big data streams
• @MrHenHan
4. 14.07.2017
SOBIS | Our software and your power for successful projects.
4
» We would like to tell you our Python journey
• Why we (a former Java only company) love Python
• How Python supports us to achieve our product visions
• How we use Python to cluster communication and detect events
» We are the new guys… so we appreciate any feedback from you
Our Talk
5. 14.07.2017
SOBIS | Our software and your power for successful projects.
5
PIRS – Project Collaboration and Communication System
Correspondence Management
• Storing project information
• Sending, storing and tracking project
correspondence
• E-Mails, Letters, etc..
• Action Item Management & Follow Up
• Out of the Box Interface to your mailbox
Document Management
• Storing all project relevant documents
• Generating Transmittal Slips
• Controlling of the distribution of drawings
• Tracking the status of drawings
• Revision control / approval workflows
6. 14.07.2017
SOBIS | Our software and your power for successful projects.
6
PIRS - Product Technology Journey
Lotus Notes
Domino (1998)
Java Web Applications
(2009)
Java meets Python
(2015)
9. 14.07.2017
SOBIS | Our software and your power for successful projects.
9
Project Communication
10. 14.07.2017
SOBIS | Our software and your power for successful projects.
10
Project Portfolio & Communication
11. 14.07.2017
SOBIS | Our software and your power for successful projects.
11
» Managed project » Manage your project data
• Folders & projects
• Correspondence & meta data
• Controlling & follow up
• Favorites & Tags
• Reports
• Filters & Full Text Search
» Search Content
• User wants to search for a specific content or
information
• Topic is already known by the user
Project Communication
12. 14.07.2017
SOBIS | Our software and your power for successful projects.
12
» Is manual classification of information still state of the art?
» Why don’t we use the core of correspondence?
» Vision: Provided content & information
• How can we implement the possibility that our application provides content
• Presented to the specific target audience
• Content wants to find its audience
Questions & Vision
16. 14.07.2017
SOBIS | Our software and your power for successful projects.
16
Task - Identify Topics, Hot Topics and Events
Order-66
PM Invoices
Topics
• Labled Cluster
Hot Topics
• Cluster growth above average
Events
• New Cluster
• Exceptional Cluster growth
Available Information
• Participants, Content, and Meta
Inform.
17. 14.07.2017
SOBIS | Our software and your power for successful projects.
17
Communication model - social stream graph
18. 14.07.2017
SOBIS | Our software and your power for successful projects.
18
Atomic Model - Social Stream Object
𝑞1
𝑟3
𝑟2
𝑟1
𝑡1
Sender Receiver(s)
19. 14.07.2017
SOBIS | Our software and your power for successful projects.
20
Cleaning and normalization
Textual similarity
Structural similarity
Other similarities
Preprocessing and Similarities
20. 14.07.2017
SOBIS | Our software and your power for successful projects.
21
Similarities èn Detail
Tf-idf based Cosine Similarity
Bit Vectors
Normalized Tag mutualities
23. 14.07.2017
SOBIS | Our software and your power for successful projects.
25
Tools of Trade
NetworkX
24. 14.07.2017
SOBIS | Our software and your power for successful projects.
26
Jupyter
Design
Impl.Test
Partial
Deploy
Workflow… Or when Research meets Engineering
25. 14.07.2017
SOBIS | Our software and your power for successful projects.
27
Java Web Application
Security
Authentification Authorization
ORM
Databases
Interfacing with existing Technologies
26. 14.07.2017
SOBIS | Our software and your power for successful projects.
28
Control API Ressource API
Processing Statemanagement
PIRSA
Machine Learning as a Service (MLaaS)
27. 14.07.2017
SOBIS | Our software and your power for successful projects.
29
Security
Consumer
Service
Interfacing
Design
Loosly
coupled
Process
Iterations
within
iterations
Experience
Privacy
Usage
scope
Consumer
Challenges
28. 14.07.2017
SOBIS | Our software and your power for successful projects.
30
Thank you for your attention – Questions?
29. 14.07.2017
SOBIS | Our software and your power for successful projects.
31
“So Long, and Thanks for All the Fish”
30. Europe – Germany
SOBIS Software GmbH
Fraunhoferstraße 8
68309 Mannheim
Tel.: +49 621 1228 5600
Middle East – Dubai
SOBIS Software GmbH
Building 16, Dubai Internet City
P O Box 390742
Dubai UAE
Tel.: +971 55 763 7314
Middle East – Saudi Arabia
SOBIS K.S.A.
Wasel: 8974, Zip code: 12214
Unit 4 – Wadi Ath Thumamah
Riyadh, Kingdom of Saudi Arabia
Telefon: +966 11 216 1180
Asia – India
SOBIS Software (India) Pvt. Ltd.
#12, Varsav Plaza, Jayamahal Main Road
Bangalore – 560 046, India
Tel.: +91 80 2354 7953
www.sobis.com
info@sobis.com
Editor's Notes
Short introduction Stephan und Henrik
SOBIS
- Founded in 1998 with more then 100 employess with offices in Mannheim, Bangalore & Riyadh
Core business: Standard Collaboration Solutions for engineering companies and projects:
Custom web applications, mobile solutions and interfaces
Before we start with our vision, just a littble bit more information about our technology
Journey started in 1998 – Who recognizes this icon
Lotus Notes
Was a great collaboration tool
A lot of big companies used for collaboration and map business processes and workflows
Rapid Application Development
Offline replication was a powerful weapon especially
2006 to 2008 we realized: IBM is gonna die and Microsoft will win the fight
so we have to switch to new technologies: Hello traditional Java web applications
Web / Mobile Applications
Introducing Java (Backend) Applications (also some .Net and C#)
Java Script Frameworks for UI implementation (e.g. ExtJS / Sencha Touch / Angular JS)
MySQL / MSSQL / Oracle Databases
Mobile WebApplications and Native Appl Development
In 2015 we also added Python to our development portfolio
We built a small team concisting of 7 people (mostly students but also )
and we will tell you why now
To understand our vision, some words about our customers and the challenges
Our customers:
Focused on engineeringer projects
Plants (Nuclear, Oil & Gas)
Civil & Construction
Water
Transportation
Challenges
High prices sensitivity
Complex technical solutions
Project duration up to 10 to 15 years
A lot of involved contractual parties: Customers. Consultants, EPCs, Suppliers, Sub-Suppliers
And also disciplines:
Management of communication is an essential part for the success of the project
In most projects, communication, documents, are stored within different locations, places and people.
This gets even worse if you think about the portfolio
Managed Project
Why do our users have to manually classify, tag and manage correspondence and project information?
We are great in managing the existing correspondence and information using meta data, but why don‘t we use the core of correspondence: message content?
Recommendation Engine
Identification of interesting and important topics for project team members
“This correspondence might be interesting for you”
Identification of domain experts / profiling
Evaluation of project and portfolio
Who knows who
Who is collaborating with customers, suppliers, …
Objective: Automated identification of domain experts & profiling
Trends & Risks
Identification of trends, topics, cluster and risks within
Single project
Project portfolio
Customer, contractor, consultant or supplier
Mood Detection (Customer is happy / not happy)
Clustering and Event Detection
Linking and referencing of independent communication and events
Classification and categorization of communication and content
Automated topic detection