Services

Delta Science

Get in touch

Digital Research, Data Science & Data Engineering

Focus, Sharpen & Fix
Your Data Processes

Depending on your specific project needs you can boost different aspects in a modular fashion using our support:

Data processes & data infrastructure

Data Engineering Life-Cycle

Effectively managing data flows, i.e. enhancing and transferring data to meet business needs is essential for all successful operations that rely on data. Data generation, storage, ingestion, transformation and serving all of these essential building blocks are increasingly managed in cloud-based technologies offering scalable, fast and high-throughput solutions. To mention some of the latest tools, technologies or cloud-services that we offer to support: Airbyte, AWS, Apache Kafka, Apache Spark, ClickHouse, Databricks, dbt, Docker, Confluent, Jinja, Preset, Snowflake.

Clean take-off

Clarify Priorities & Start Exploring

Each project is loaded with a lot of unknowns. Ideas, requirements, opinions and interests all with varying degrees of empirical foundation (including zero). Clearly, it's hardly feasible nor efficient to explore everything. However, how do we know which areas are absolutely crucial to get more certainty about our assumptions to increase the success of a given project? To select the right starting point for exploration we need to identify and rank risks that can derail the envisioned project. The resulting area is the most important one to start separating facts from assumptions.

Sharpening the focus

Drill To The Core For Testing

It is essential to weed out any unnecessary questions in order to identify the most important one. From all the possible questions in the selected area remove all clutter and drill down to the most fundamental one. This core question needs to be transformed into a testable assumption (hypothesis), i.e. an assumption that can be proven wrong in the face of data, i.e. in the presence of empirical evidence. This is the only way that the assumption can be effectively tested in an experiment.

Real-World evidence

Solid Experiments: Getting Actionable Data

Shaping solid experiments to gain high-quality data means to establish a controlled framework for data collection in precise reference to your hypothesis. Experimental design includes a range of different steps, whose exact composition depends on the nature of the investigation: e.g. defining test condition/s, inclusion/exclusion criteria for participants, technical setup and quality control, data scope/points, outcome measure/s, statistical power/sample size estimation, data analytics strategy, statistical methodology, minimum success criteria, data collection monitoring; all-together contributing to a flawless and controlled setup.

Digging for treasures

Data Science: A Craft Not A Magic Blackbox

Data Science is a vast field of diverging disciplines, activities and methodological approaches. It's easy to get lost in terminologies and overwhelmed by latest trends in AI/ML. Our approach is guided by one a basic principle: using Data Science as a craft that delivers hands-on solutions rather than a black-box marketed as pure "magic". Therefore, developing and providing commented source-code for data analytics pipelines (transform, visualise, model) is for us an essential commitment to transparency as well to reproducibility of results and scientific integrity. (execution: Python, R, SQL , Matlab incl. div. packages)

Bridging the gap

Integrating Qualitative & Quantitative Data

Learning from customers or users may entail not only quantitative, but also qualitative data. However, too often project teams shay away from collecting customer responses in real conversations, possibly due to the more cumbersome and delicate way of collecting and analysing this type of data. In addition, quantitative data are omnipresent due to the broad use of digital environments and the traces of user activity become readily available. Depending on the nature of the investigation, it might become mandatory to understand not only WHAT users are doing, but WHY they are doing it. In that case qualitative data via conversation with real people becomes your friend. Done right, i.e. avoiding the mistake of just surveying opinions, this road provides a unique and complementary picture on user behaviour: the WHY behind the WHAT.

Accessible insights

Data Analysis, Code-Based Visualisation & Reporting

The key objective is to create compelling graphs that not only communicate the insights extracted from the data effectively, but also direct the reader's attention seamlessly in sync with the conveyed meaning. Many different professions, ranging from research scientists to communication specialists are occupied regularly with this task using high-level expertise to solve it. However, some projects additionally need dynamic visualisation routines that can be updated based on the availability of new data, implementing the pre-defined visualisation principles and matching all graphical style details in a precise and reproducible manner. We support to establish code-based visualisation that includes data pre-processing and data cleaning procedures that can be used in dynamic reports.

Data anywhere, anytime

Make Analytics Tools Work: Leverage What Is Already In Place

Most companies or projects have one or several web/app analytics tools integrated in their digital environments tracking user visits on different levels of granularity, depending on the technical characteristics of the tool, business needs, regulatory restrictions and individual setup adjustments. Popular analytics tools are GA, Adobe Analytics, Mixpanel, Matomo, Countly, etc - just to name a few. Whenever you want to look beyond the basics of user visits it's worth investigating what treasures you can realise within the established tracking infrastructure. Tracking tools have become very powerful and flexible tools, but at the same time their feature-richness can provide a real obstacle to use them properly and to make the necessary adjustments in their settings. We can help you to find your optimum and get what you need by leveraging the capabilities of the tools in place (rather than getting new ones).

Statistics & Data Science: Solid Data-Based Decisions

Solid protection

Fight Randomness To Enable Solid Data-Based Decisions

Nobody wants to make a wrong decision. However, being confronted with random results this threat becomes reality. To base a decision on incidental findings means to derail the whole insight generation process. The ultimate goal of any exploration is to arrive at a point to draw a solid conclusion based on the data at hand. That is, to make a data-driven decision. As a matter of fact, you want to be confident to make the right decision having some level of trust that your results are in fact real and not random. Statistics done right helps you to asses the confidence you can have in a given result to protect your decision making. However, the realm of statistics is for far too many a pretty uncharted territory. We can help you navigate your project based on established best practice methods taking into account classical statical tests as well as modern approaches of permutation statistics.

Pick the support you need

Typical Project Formats

To discuss your specific project scope and needs we should have a chat. Together we will define the right format of support. To give you a first impression here is a list of most common project formats:

Small

Step-By-Step

Do you want to start on a clear-cut challenge? We can start small supporting you in one specific task, for example a validation experiment, analytics setup adaptation or data analysis task.

Extended

Full-Fledged

In this format we work hand-in-hand with your team focused on getting the insight generation process running, in line with the overall-project aim and reinforcing the necessary skillset and empowering team members.

Big

Network-Size

If your project is large and requires assistance in multiple areas, including digital product development (mobile & web), project management, strategy and business planning, financial operations, and personnel organisation, I have access to experienced professionals who have already demonstrated their competence and ingenuity in a variety of past and current projects.