Our Strengths

Our Strengths

What makes us special

What makes us special

We combine the methodologies behind Machine Learning and classical Statistics with insights from behavioral sciences to allow for a fusion of human know-how and technological advancement.

Working with us holds some unique advantages for you

flash_on

External Data

We have developed a cost-effective means to find and integrate relevant external data for any use case. Since more (relevant) data generally leads to more precise predictions, we will offer you some of the most precise models in the market.

build

Expertise

We are familiar with the most recent developments in Big Data and Machine Learning as well as the current industry standards in either field. Additionally, we can offer our Cyber Security Expertise, which allows us to secure solutions developed for you.

business_center

Crowdsourcing

We bring hundreds thousands of freelancers with knowledge in all flavors of digital work at your finger tips. Deploying our crowdsourcing toolset will also allow you to involve your employees in our prototype- and solution development

External Data

External Data

Use our toolset to discover, acquire and integrate relevant external data

Use our toolset to discover, acquire and integrate relevant external data

Data is the currency of the digital age. Whether it’s data about your customers or data to better understand your competitors: data is key to any analytics! Internal data is often easily usable, but in many cases not enough to get a well-rounded picture of the goal-agent (e.g. your customer). Therefore, companies in need of good analytics models spend a significant share of data science projects finding the right external data sources to complement the internal data.

There are a few major pitfalls involved with finding the right external data sources, where data science experts currently struggle and spend a lot of resources for.

One of the key assets of the founding team of PeakData are the results of their research, where they have developed tools for just this problem. We have developed a means to:

  • Deploy crowdsourcing to discover what data sources could be relevant
  • Obtain this same data for you - at higher result quality than using in-house resources and at lower price
  • Model both the internal and external data and deploy it into production

We offer the DOM services as full projects, as well as support to existing data science teams.

Expertise

Expertise

Use our knowledge to your advantage

Use our knowledge to your advantage

Our employees cover the full skillset required for Big- and Small Data projects. Our employees are certified Amazon Cloud experts, which can enormously reduce cost (and increase throughput) of analysis projects. Due to having Cyber Security education in-house, we are well-positioned to further reduce (data) security concerns of the cloud.

Classification
  • Deep-Learning / Neural Networks
  • Generative Adversarial Networks (GAN)
  • Support-Vector Machines
  • Random Forests
  • Decision Trees and stubs
  • Rule Learners
  • [..]
Clustering
  • DB Scan
  • K-Means
  • Hierarchical Clustering
  • [..]
Data Types
  • Time Series Data
  • Unstructured Textual Data
  • structured Textual Data (XML, JSON, etc)
  • Spatial Data(Geo Data)
  • Graphical data
  • Streaming Data
  • Graph Data
  • [..]
Visualization Tools
  • GGPlot
  • OpenLayers (for Geodata)
  • Tableau
  • Shiny
  • [..]
Big Data Tools
  • Amazon AWS, EC2, Lambda
  • Amazon EMR, Kinesis
  • Apache Spark, Hadoop, Kafka etc
  • [..]
Storage Technologies
  • Archives (CSV, JSON etc)
  • SQL (+flavors: MySQL, Oracle, Postgres, Amazon RDS, ..)
  • NoSQL (Mongo, Cassandra, Amazon DynamoDB, ..)
  • Neo4J (graph data)
  • Apache HDFS
  • [..]
Languages
  • Java 7-9, Scala
  • Python 2 and Python 3
  • R
  • PHP, HTML, CSS, Javascript, ...
  • [..]
Crowdsourcing
  • Open-Call settings and precise worker selection
  • Automatisches verwalten von Lang- und kurzlebige Arbeitsbeziehungen
  • Programming, Data Science, Data Visualization and Data Preprocessing tasks
  • [..]

Crowdsourcing

Crowdsourcing

Deploy thousands of freelancers in all digital disciplines at your finger tips

Deploy thousands of freelancers in all digital disciplines at your finger tips

Crowdsourcing is an emerging technology that allows algorithms to interface with human freelancers distributed across the globe as well as employees within your organization. We use crowdsourcing in three major ways:

  1. To ideate with your employees
  2. To reduce cost in Data Science projects
  3. To improve the prototypes we develop for you

Crowdsourcing for Open Innovation

Open Innovation is one of the best-known applications for Crowdsourcing: it provides a platform for a large number people to come up with ideas for a given question together.

In recent years many companies drive innovation by means of Open Innovation tools.

We at first applied it for a major question in data science: how can data science help the different functions of a specific company.

To unlock the potential of data, we enable employees across the organization to ideate on different opportunities for data-driven methods they recognize in their field of expertise. This deliberation is coupled with expert knowledge – depending on your security vs cost needs coming from PeakData or a number of third party freelancers.

Eventually. employee ideas are filtered and ranked according to their feasibility and payoffs. The best ideas are then brought to your attention.

Bonus: Besides employees throughout your organization, we can seed idea generation through hundreds of third party freelancers.

Crowdsourcing for Data preprocessing

Data preprocessing is among the most time intensive tasks in Data Science projects. It is well known, that 80% of the time in Data Science projects is spent on Data Preprocessing. Saving time in Data Preprocessing therefore directly translates to large cost savings in data science projects.

Before we founded PeakData, we have researched, built and tested means to deploy both external and intra-organizational crowdsourcing for data preprocessing.

  • External crowdsourcing: recruitment of crowds across the world to deal with data transformation required for further data analysis. This scenario is suitable when the privacy is not an issue.
  • Intra-organizational crowdsourcing: employment of employees in your organization who possess some basic coding skills in order to transform data. You can imagine a scenario where interns and junior software developers working together with data scientists. This scenario is suitable when the privacy is important.

Using crowdsourcing instead of internal resources for data preprocessing yields the same quality output at roughly half the cost and duration – and it keeps Data Scientists happy, since nobody really likes (data) cleaning.

Crowdsourcing for prototype/solution improvement

Did you know that 30%* of software products fail? Wonder why? requirements engineering!

You can choose to have the prototypes we ship, as well as the solutions developed for you, contain crowdsourcing components built-in.

These components allow you to easily engage with the users of a solution/prototype and collect/aggregate their feedback throughout its lifecycle This ensures, that end-users remain in the requirements engineering and development process of solutions and therefore guarantees a higher project success rate than in classical settings