Monthly Archives

January 2019

Confusion Matrix showing True Positives, True Negatives, False Positives and False Negatives

False Negatives: Evaluating Impact in Machine Learning

By | AI & ML | No Comments

Recently, I had the opportunity to build a regression model for one of FTS Data & AI‘s customers in the medical domain. Medical data poses an interesting challenge for machine learning experiments. In most cases when running algorithms for binary classification, the expected result in the training set will contain a large percentage of negatives. For example the goal of an experiment might be to predict if – based on a set of known clinical test results – a patient has a certain medical condition. The percentage of positive results in such a set, if it is a generic dataset for a vast number of medical conditions will most likely be very low. As a result a machine learning model when initially tested using a small set of chosen features will most likely come up with a high number of false negatives.

The latter however is a big problem in experiments involving clinical data, i.e. categorising that a patient does not have a certain medical condition incorrectly could have disastrous consequences. Once a confusion matrix is built, the model’s effectiveness is measured using indicators such as area under curve, accuracy, precision, recall and F1 score. In medical datasets, recall plays a big role as it measures the impact of false negatives. It can therefore hold significant weight in determining the most appropriate model for a given experiment.

The definition of recall is –

Recall = (True Positives) / (True Positives + False Negatives)

In the confusion matrix, the denominator in this equation makes up the total actual positives. So, recall therefore is effectively measuring the correct positive predictions over the actual number of positives in the dataset. If there were no false negatives, recall would be at the ideal score of 1, however if a large number of actual positives were predicated as negatives (i.e. false negatives), recall would be much lower.

As the model evolves and more relevant features are chosen for prediction, recall should start improving. In domains such as medicine where false negative predictions can have dire consequences, the recall score should play a vital role in choosing the most optimum model.

Microsoft Gold Partners Data Analytics

By | Uncategorised | No Comments

We are excited to announce that FTS Data & AI have now become Microsoft Gold Partners in Data Analytics! This is an important development in our ongoing partnership with Microsoft in Australia. It shows we have reached the high standards expected of partners that qualify for the program. The Data Analytics competency is one that shows we have a deep expertise in the associated technologies – SQL Server, PowerBI and other Azure capabilities such as Azure Data Factory.

FTS Data & AI are Microsoft Gold Partners Data Analytics

Microsoft Gold Partners Data Analytics

Microsoft are a key partnership for us as they provide an industry leading platform for delivering Data Analytics solutions to our customers. Us achieving this partnership level is a demonstration of the capabilities of our technical team. It is also a recognition of the customer successes we have helped drive.

What does Microsoft Gold Partners in Data Analytics mean for customers?

Becoming Microsoft Gold Partners in Data Analytics has customer benefits in terms of confidence in our expertise and access to the latest technology.

For our customers it demonstrates that they are working with a technology partner that has certified personnel, with access to the latest in training and product updates. We have to maintain a number of certified people on our team who are qualified in the Data Analytics platforms that Microsoft provide. These certifications cover on premise and cloud technologies and range from traditional Business Intelligence through to Big Data solutions. Microsoft Gold Partners also get access to cutting edge product training as we are encouraged to upskill on the latest platforms.

Gold Partners also benefit from product information & roadmaps and industry insights direct from the Microsoft Data Platform team. As the product suite is updated we receive access to product news and information. We are also given Azure Credits so we can trial the new capabilities as part of our commitment to keeping our customers informed and aware of the latest developments.

D365 FinOps Reporting options

By | Data Visualisation | No Comments

The D365 FinOps Reporting landscape is a bit tricky to navigate. The documentation has not keeping up with the pace of product development. This makes the technical complexities difficult to navigate. In this quick post we provide an overview of what options you have. Dynamics 365 Finance and Operations (aka D365 FinOps) is a powerful ERP that can significantly improve business process efficiency, but for us at FTS Data & AI, we also see that there is a lot of additional value to be obtained from the data it captures.

What are the D365 FinOps Reporting options?

The three key tools within the D365 FinOps Reporting suite are:

  • Financial Reporting (aka: FRW)
  • PowerBI
  • SQL Server Reporting Services or SSRS (aka: Document Reporting Services)

There is also a fourth way – which is to pull the data out (aka BYOD or Bring Your Own Database) and report on it with whatever you like, but that’s for another post.

Financial Reporting

The simplest option – Financial Reporting (previously Management Reporting) – has a designer built into the D365 environment. It’s purpose is to provide Management Financial Reporting – i.e, Balance Sheet, Cash Flow, P&L – with some ability to customise your financial reports to your organisations viewpoint, with some filtering available for operational segments.

However its key limitation is that it is Financial in nature only and geared towards providing fairly static, statutory type reporting. Add to that the design and build of reports is very manual and it quickly reaches its limits. A typical example of such a report is shown below – you can see it is very much geared towards line item summaries.

D365 Finance Report Writer

D365 Finance Report Writer

PowerBI

PowerBI is planned to be the key reporting tool for D365 and may well replace the other options over time. A significant number of the embedded reports are already built in PowerBI. Part of any D365 Subscription is a PowerBI Embedded instance. All of the PowerBI reports in D365 are managed through this. Typically the generic reports do not suit the individual business needs and thus they need customising or extending. This is possible – complex to get started – but simple enough to execute once initiated, for example leveraging our Data Visualisation team who have deep experience doing this. Below is a “Before & After” set of examples to show the difference between out of the box reports and ones that have been enhanced by our team.

D365 FinOps PowerBI Report

D365 FinOps PowerBI Report – Before

 

D365 FinOps PowerBI Report

D365 FinOps PowerBI Report – After

A key restriction is that the embedded reports have to run off of the data in the Entity Store. This is a subset of the data in D365, with predefined and scheduled aggregations. To add content to the Entity Store requires a D365 developer, which adds cost to the report development process.

SQL Server Reporting Services (SSRS)

In every module in D365 there is a substantial set of SSRS reports that are included under the name of “Document Reporting Services“. These are embedded into D365 and present views of data with very simple parameters. You can view the technical details of included reports here and there are at least a thousand of them.

As with the PowerBI option it is possible to customise, brand and extend these reports. In this case the data source is any query that can be written against the Application Object Tree (AOT). Importantly this can show transactional as well as aggregate views. This allows for the production of customer facing content such as invoices.

Summary

So there are three options, each with their quirks:

  • FRW for detailed Financial Reporting
  • PowerBI for Aggregate Analytical Reporting
  • SSRS for highly customised reports

It’s not the clearest or most well documented landscape to navigate, but it is generally possible to produce the content you need to meet your organisations reporting requirements.

PowerBI Partner Showcase is published

By | Data Visualisation | No Comments

We are excited to announce the publication of our first PowerBI Partner Showcase! The Government Contract Analysis Tool is a demonstration of our ability to deliver powerful data visualisation solutions. It’s a great piece of work by our team and a big recognition of their capability.

Power BI Partner Showcase - Government Contract Analysis Tool

Power BI Partner Showcase – Government Contract Analysis Tool

 

Based on Open Data this PowerBI Partner Showcase demonstrates our ability to visualise data to provide insight on Federal Government Spending patterns – check it out on the official PowerBI site. It’s interesting to see how many contracts are awarded in June. This is visible in the bottom chart of the first page. In 2002 a whopping 83% of contract were awarded in June. Surprisingly, not an election year.

The Government Contract Analysis Tool

So what purpose does it serve? Government contracting is a vast industry worth billions of dollars with thousands of contracts awarded every year. Such large data presents difficulties in offering useful analysis, resulting in the commercial trends going unnoticed.

Typical questions are:

  • Who are my major competitors for this service?
  • What procurement method should I tender with?
  • When is the best time to tender for work?
  • Where are the most contracts being awarded within Australia?
  • How long will I need to commit for a contract with this agency?
  • What growth has there been for contract values across agencies over time?
  • How much should I tender for?

To remain competitive and drive revenue growth, decision makers need accurate and real-time answers to the above questions. Unfortunately, direct and constructive analysis is difficult with the vast quantity and low quality of the data provided.

FTS Data & AI took the raw data and crafted an effective analysis tool that provides answers to the critical questions that decision makers demand.

Users can now:

  • Identify who are their major competitors, and how much they have earned in contract revenue, thus helping shape price and value propositions.
  • Determine what procurement method is preferred for each service and agency, reallocating resources and refining tender strategy to better suit the favoured method.
  • See when most contracts are awarded in a financial year, and how much the average value is for that contract, allowing ample time to plan executable strategy.
  • Locate where most contracts are being awarded in Australia, enabling users to lift and shift resources into areas with greater potential for work.
  • Track how long different agencies and services are demanding contracts span, giving users the ability to budget time and associated costs of work.
  • Recognise trends in growth of contract value for each agency and service, empowering decisions around service offerings and prices.
  • Monitor market prices for each agency and service, sliced across numerous fields, informing users of various price pressures and trends.

Using Power BI, FTS Data & AI has transformed a low-quality dataset into a fully-interactive reporting tool capable of providing clear and concise answers to the questions that decision makers are tasked with solving to ensure continued business success.