Data Science

Harnessing The Power Of Data

Axiologic Solutions doesn’t just “Do” data science – we live it and breathe it, through every project challenge including cyber security threats and identifying bias. Our range of professional services cover every step through the full data science lifecycle, including:

  • Problem statement creation
  • Hypothesis creation
  • Data sourcing
  • Data engineering
  • Data wrangling
  • Data visualization
  • Analytical technique selection
  • Data analysis
  • Analytical model creation
  • Model evaluation (including cyber security threats and bias)
  • Model packaging, model integration, and model performance monitoring

Axiologic Solutions

Agile Data Science Methodology - disENTROPY

Our data science lifecycle is encoded in the Axiologic Solutions Agile Data Science Methodology – disENTROPY™, which our data scientists use to ensure maturity and consistency in how they perform their work.


The Axiologic Agile Data Science Methodology

Our data science methodology includes four components – data modality, analytical techniques, applications, technologies/tools

Data Modality.

We have expertise with following data types/modalities:

  • Natural language text e.g., English, Spanish
  • Transactional data: various domains, including scientific data
  • Financial data: various domains; various software systems (e.g., Momentum®)
  • Geospatial data
  • Imagery
  • Social media data
  • Open-source data
  • IoT device / Sensor data: various sensors (e.g., SIGINT RF)
  • Graph / network data

Analytical techniques.

We employ three main classes of techniques to perform data analysis:

  • Mathematical Techniques:
    • Calculus (derivatives, differential equations), geometry and linear algebra
    • Vector, matrices, and Markov Chains
  • Statistical and Probabilistic Techniques:
    • Regression analysis (linear, non-linear)
    • Analysis of variance (ANOVA)
    • Correlation analysis
    • Cluster analysis
    • Uni/Multi variate analysis
    • Discriminant analysis
    • Factor analysis
    • Time series analysis
    • Bayesian analysis
  • Machine Learning Techniques:
  • First Generation Machine Learning
    • Linear Regression
    • Clustering (various)
    • Naive Bayes
    • Decision Trees and Random Forests
    • Rules Extraction
  • Second Generation Machine Learning
    • Logistic Regression
    • Auto-encoders (various types)
    • Support Vector Machines
    • Conditional Random Fields
    • Hidden Markov Models
    • Feed Forward Artificial Neural Network (multi-layer perceptron)
    • Deep Neural Network: CNN, LSTM, ESN, RNN, VAE, GAN
    • Gradient Boosting
  • Third Generation Machine Learning
    • Embeddings
    • Transformers
    • Attention / Memory Learning
    • Sequence Network
    • Capsule Learning
    • Graph/Network Learning
  • Cutting edge
    • GPT-3.5/4
    • Gopher
    • LaMDA
    • DALL-E
    • Embedding (xToVec: Graph, Structure, Node, Doc, Sentence, Word, etc)
  • Miscellaneous Techniques:
    • Fuzzy logic
    • Formal logic (Propositional, etc.)
    • Rules-based models
    • Data mining techniques
    • Monte Carlo simulation


We have expertise in the following well-known (and some lesser known) applications of ML models:

  • Natural Language Processing (NLP), including:
    • General (tokenization, NER, POS, word sense disambiguation, coreference resolution)
    • Topic Modeling (or Topic Extraction)
    • Document Summarization: single and multiple
    • Stylometry Analysis / Authorship Detection
    • Virtual Assistant / Chatbots
    • Document Similarity
    • Spelling and grammar correction
    • Sentiment Analysis
  • Computer vision, including object detection, scene understanding, text detection, and object segmentation
  • ML for Cyber Security
  • ML for Insider Threat Detection
  • ML usages in Robot Process Automation (RPA)
  • Recommenders (various)
  • ML for graphs/networks


Our data scientists have experience with the following tools and technologies:

  • Programming languages: R, Python, Java, Scala
  • ML libraries:
    • Spark, Python, or various open-source projects
    • scikit-learn
    • TensorFlow Extended
    • Keras
    • Caffe
    • Theano
    • Networkx
  • Workflow tools: Kubeflow, Apache Airflow, Apache NiFi
  • Cognitive pre-packaged services
    • Azure ML tools (various)
    • AWS ML services (various)