DATA SCIENCE TECHNIQUES
To truly understand your business it is critical to model the behaviour of its component parts. A core concept behind the concept of data modelling a business’s services and products is iteration. Every time an iteration of the data model is complete it is essential to stop and assess the accuracy of the model in reporting on and predicting actual measured behaviour in the real world. This assessment is then used to refine the next iteration of the model, often taking into account unexpected emergent insight into your business.
Pebblewash is experienced in the iterative process of creating and refining business data models, and in adapting and incorporating additional unexpected insights as they arise. The key to modelling a business accurately is to understand there is always another layer of knowledge waiting to be unwrapped
At the bleeding-edge of data science lies the concept of machine learning where a system is designed to teach itself about the co-relations of the dataset it works with every day, even as the contents of the data set grow and change. Machine learning is focused on the whats of causation, not the whys or the hows sharing insights that would never be reachable through the application of common sense or standard statistical analysis of representative data subsets.
Pebblewash brings an in-depth understanding of how to design, deploy, and refine machine learning installations to glean the utmost insights from the largest data sets. We can help you build a self-learning system that never stops digging for the next unexpected revelation.
In the world of data science, graph theory encompasses a number of inter-related techniques that provide powerful insight from data sets composed of nothing more complex than people’s views and interests casually expressed on the internet.
Pebblewash’s experience with applying graph theory to combinations of taste profiles, social networking, sentiment analysis, and recommendation engines ensures we can help your organization delve deeply into the relevant data to extract additional value and insights for your company.
Time series analysis and Prediction
Time is an integral element in the useful analysis of big data, extracting additional insight by measuring against what happened in the past, what’s happening right now, and what’s likely to happen in the future.
The Pebblewash team has extensive experience using time series analysis and prediction to offer guidance on what a company should expect to see and when.
Map reduce is a massively parallel processing distributed computing technique most familiar from Hadoop but employed by a number of tools in the big data and data science toolset. Map Reduce is a key approach for managing data sets containing billions of individual records, cleverly distributing the computational load across the total number of machines in your MPP cluster.
Pebblewash is familiar with all the inner workings of the Map Reduce approach and has used Map Reduce with several MPP databases.
Network Failure Prediction
In the information age lives in a world of networks, and networks are the life blood of many organizations, often existing in complex symbiotic interrelationships with other networks, both those of strategic partners and others providing the skeletons of regional ISPs or the very backbone of the internet. For any company that lives and dies by their network, there is nothing more critical than constantly improving their understanding of their network.
Pebblewash’s team will help your NOC model, monitor and predict the health of your network and the networks your network relies on.