Assessment of your initial situation and your individual requirements for an analysis infrastructure in a telephone interview.
eoda | analytic infrastructure consulting
Best practice approach for the professional application of data science
Data science has become a crucial success factors in many industries. Programming languages such as R, Python or Julia are the key to this success factor. In order to tap the potential of data science lastingly, a seamless implementation of analytical solutions into business processes is required.
This is where our eoda | analytic infrastructure consulting comes into play. It is the best practice approach for the professional operation of your data science architecture. Version management, rights management, documentation or testing – we combine the most important components so that you can implement analytical processes sustainably in your business.
From the daily work with the programming languages themselves to the linking to big data technologies such as Apache Spark or Hadoop to the connection to the existing IT environment – make the right tool and technology decisions with eoda | analytic infrastructure consulting and create the optimal analysis environment for your requirements.
The benefits of the eoda | analytic infrastructure consulting for you:
Reliability: We combine proven best practice approaches to create security for the professional operation of your analytical environment.
State of the art: Data science teams can achieve maximum performance with the recommended state-of-the-art techniques and processes.
Process validity: Based on our consulting you can create an analysis environment that enables you to generate valid processes by means of role definition, documentation and version management.
Process efficiency: Integrate your analytic stack perfectly into your data-driven business processes.
Collaboration: With our support you will be able to facilitate the collaboration between data science and business departments as well as the onboarding of new staff and external service providers.
Professional partner: Benefit from the experience and know-how of the data science enabler eoda.
Gain a comprehensive overview of the status quo of your analytics environment and receive a detailed vendor-neutral realization concept with roadmap and precise recommendations for the professional application of data science in your company.
One-day inhouse workshop in which the current situation will be analyzed and existing potentials will be identified.
Detailed report with an evaluation of the status quo and an individual roadmap for optimizing your data-driven processes.
After the assessment you can implement your analysis environment - with the professional support of eoda or on your own.
The architecture has decisive effects on performance, agility and administration but also on the individual access rights of users.
Examples: Client Server, desktop, batch or streaming architectures
Automation + Deployment
Infrastructure for event and time-controlled execution of analyses and automated testing.
Examples: Continuous Integration Scenarios with Jenkins
Data Science Languages
Which programming language is best suited for the requirements in your individual analysis scenario?
Examples: R, Python, Julia
Integrated Development Environment
From file management to syntax highlighting - the decision for a development environment significantly affects the usability of R.
Examples: RStudio, Eclipse, Visual Studio, PyCharm, Jupyter, Spyder
Unlike Python and Julia for example, there is a whole range of interpreters for R. They differ especially in performance, e.g. with regard to the analysis of big data.
Examples: GNU R, Oracle R, Oracle R Enterprise, Microsoft R, TIBCO TERR
The productive application of programming languages such as R and Python imposes special requirements on the use of packages in terms of managing dependencies or deploying critical updates.
Examples: miniCRAN, Packrat, checkpoint, pip, Anaconda, Miniconda
Depiction of workflows with individual role definitions and a granular authorization concept.
Example: Integration LDAP System, Integration Active Directory
Static PDF, HTML document or interactive web application - implement the optimal toolset for your reporting.
Create different environments for different maturity levels of analyses - from the development to the live environment. Implement processes for the handling of release changes and software updates.
Examples: Dev, Test, Live
Systematic testing of code in order to identify and lastingly eliminate errors. In this way you will increase the quality of the analyses on which your business decisions are based.
Examples: testthat, nose, unittest, individual tests
For the professional application of analysis scripts it is recommended to use version management in order to record changes to the source code comprehensibly and transparently.
Examples: Git, Git-Flow Branching Model, GitLab, GitLab Enterprise, Atlassian Bitbucket, SVN, TFS, Semantic Versioning
External Analysis Frameworks
For specific analysis scenarios such as Deep Learning or linear optimization there are external frameworks which can be addressed via APIs. We evaluate and implement the appropriate framework for you.
Examples: MXNet, H2O, IpSolve, TensorFlow
The correct documentation of code or the definition of a consistent style guide facilitates the collaboration and efficiency of the analysis team as well as the reproducibility of the processes.
Examples: Roxygen, packages, wikis, Sphinx
The implemented framework can be operated by the eoda analytic infrastructure team or by your own IT.
- Monitoring of server and software components
- Evaluation of logfiles
- Observation of the release plan of components with regard to critical updates
- Creating new users or extending storage capacity
- Rapid fault rectification
- Regular as well as unscheduled updates of all components of the analysis environment
- Support of users and IT in case of problems with the operation
- In addition: assistance of data scientists for questions regarding R, Python or Julia