Sign in

Proactive listening to customer feedback is important to improve the experience of mobile apps. Now a day, rely on apps review rating are not enough, comment from reviews are also important to understand what is going wrong on experience. Thanks to NLP (Natural Language Processing) technology is becoming more mature, we can easily understand customer feedback automatically and perform advanced analysis to continuously improve our mobile apps.

Microsoft Azure Text Analytics API is a cloud-based service that provides Natural Language Processing (NLP) features for text mining and text analysis, including sentiment analysis, text summarization, key phrase extraction, language detection, and…


In past story, I shared about how to implement logging for Java Spring Boot Containerized App and feed it into Log Analytics (Container Insight). This story is part 2, I will share about how to perform logs analytics from Azure Databricks. Below is high-level architecture diagram of environment setup.

  1. Log Analytics (Container Insight) stored logs of various containers from AKS. And administrators or developers are able to perform log queries from workspace. This part of configurations has been shared and done in past story.
  2. Log Analytics workspace export log data into Blob Storage in hourly basis.

2.1 Azure Data Factory


I haven’t shared story about Application Containerization and Azure Kubernetes Service (AKS) for a while already. Recently, I’m working with development teammates and providing advisory about logging practices for containerized apps (Java Spring Boot) which running inside AKS.

AKS have out-of-the-box containers logging and monitoring integration with Azure Monitor (Container Insight with Log Analytics). Container Insights used a containerized version of the Log Analytics agent (aka OMS agent) and running inside AKS kube-system namespace to continuously capture container metrics and logs into Log Analytics workspace.

Reference flow diagram of Container Insight.

But, by default, Container Insight is only collecting stdout & stderr from containers in AKS…


I joined Pneumonia Classification Challenge in Chest X-Rays (CXRs) recently. This round of machine learning challenge is organised by DPhi Community & Segmind. As you maybe aware, “Pneumonia killed more than 808,000 children under the age of 5 in 2017, accounting for 15% of all deaths of children under 5 years. People at-risk for pneumonia also include adults over the age of 65 and people with preexisting health problems.” - WHO

While prevalent, diagnosing pneumonia in a CXR accurately is difficult. Expert radiologists are required to review the CXR and also require confirmation through clinical examinations. …


A microexpression is a facial expression that only lasts for a short moment. It is the innate result of a voluntary and an involuntary emotional response occurring simultaneously and conflicting with one another.

Recently, I’m working on interesting project to capture and analyze peoples microexpression in real-time. In old days, we can use OpenCV + DLib + shape_predictor_68_face_landmarks model to achieve. But how to achieve it with cloud based AI models in era of cloud computing. The Azure Face Service provides AI algorithms that detect, recognize, and analyze human faces in images. Under face detection, it provided Face Landmarks and…


Happy New Year! Hope everyone have a great and healthy year. First of January is just a new day, new day with new challenges. Recently, I’m working on data engineering, science and machine learning platform with Azure Databricks. One interesting challenge is about the authentication methodology from Databricks Notebook (Python) to Azure SQL Database.

From of official documentation, the most easiest way is using PySpark with SQL User ID and Password for the authentication via JDBC driver. Azure SQL Database is also supported Azure AD (Users, Groups, Service Principles and Managed Identities) based authentication. In Azure Databricks, most straightforward way…


Healthcare is becoming most important industry under currently COVID-19 situation. Machine Learning can help healthcare industry in various area, e.g. Medical Imaging Diagnosis, Identifying Diseases and Diagnosis, Epidemics Outbreak Prediction, Medicine R&D, etc… There are tons of interestingly valuable data, e.g. X-Way/CT/MRI/PET image, medical records, etc… to feed our data hungry machine and deep learning models, e.g. CNN (Convolutional Neural Network).

Recently, I’m started to explore on machine learning for Medical Imaging Diagnosis. CNN would be great for medical image diagnosis. In this story, I tried to build simple CNN model and classify Chest X-Ray image to identity is it…


Recently, due to job related, I’m helping my customer to build data science and machine learning platform solution on Azure. One interesting and trending challenge which they want to solve is how to serve Tensorflow models easily and faster.

To expedite machine learning models delivery, it’s not limited to model serve. It’s required optimization and streamline end-to-end process flow of entire machine learning development lifecycle. Similar as application development, DevOps can be also applied into machine learning development, MLOps (Machine Learning DevOps).

In Azure, you can use Azure Machine Learning service and Azure DevOps to build end-to-end MLOps workflow. From…


Time is going fast. Few years ago, I shared first machine learning story about insurance claim prediction. It’s based on python code with logistic regression algorithm to build simple classification model as demonstration purpose.

In 2020, it should be the year of Automatic Machine Learning (Auto ML) to make machine learning process clean, simple, fast and everyone can taste it, even peoples haven’t knowledge in machine learning or data science.

Recently, due to job related, I’m helping my customer to explore/evaluate data science and machine learning platform solution. …


Believe you’re well aware of there are Coronavirus (2019-nCov) outbreak from Wuhan, China recently. Comparing with SARS from China in 2003, communication technologies are more matured now a day. There are so many datasets being published and opened from various parties. Big Data analytics are playing important role to locate, track, analyse infected persons, e.g. location being infected, path, etc…

I’m trying to make a simile map visualization for Wuhan Coronavirus (2019-nCov) cases in Hong Kong by GeoPy (Google Map API) & Folium as demonstration. The source data is from Centre for Health Protection Hong Kong (CHP) open dataset of…

Eason

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store