Microsoft Azure Machine Learning Studio Reviews

The first time that I used this tool was in a project related to bike usage in the city of Boston. This project was part of a course that I concluded some months ago. In this project I used components to read data, for exploratory analysis, for steps of data munging, to split data, select hyperparameters, and some machine learning algorithms. In some steps I needed to insert R modules to apply some data transformation.

The target of this exercise was to predict bike usage in a day.

With this tool we could have all benefits of a cloud environment, such as scalability and access to machine-learning applications. These features are very important when you have large datasets and critical applications.

• It is very easy to test different kinds of machine-learning algorithms with different parameters. You choose the algorithm, drag and drop to the workspace, and plug the dataset into this component.
• When you import the dataset you can see the data distribution easily with graphics and statistical measures.
• Easy to deploy and provide the project like a service.

For my project/exercise, this tools was perfect. I would like to see modules to handle Deep Learning frameworks.

No issues with stability.

No issues with scalability.

I didn’t need to use the support, but this tool has great documentation.

Nowadays I use Python (Anaconda and Jupyter Notebook) and R (RStudio) to create my solutions and machine-learning models.

It was very simple and straightforward. It is really simple to start building a project.

There are two kinds of licenses, Free and Standard.

Free

• 100 modules per experiment.
• 1 hour per experiment.
• 10GB storage space.
• Single Node Execution/Performance.

Standard – $9.99/seat/month (probably a data scientist)

• $1 per Studio Experimentation Hour. You will pay according to the number of hours your experiments run.
• Unlimited modules per experiment.
• Up to seven days per experiment, 24 hours per module.
• Unlimited BYO storage space.
• On-premises SQL data processing.
• Multiple Nodes Execution/Performance.
• Production Web API.
• SLA.

You will be able to create your machine-learning project and extract insights from it just by dragging and dropping components and adjusting some parameters. This tool is very user-friendly, so without a lot of programming skills you can build machine-learning projects. 

If you need more control over machine-learning modules you will need to add R or Python modules to create a customized machine-learning model.

My  primary use of ML Studio is to experiment with different algorithms and learn the techniques of machine learning. In the meantime, I have developed a few models related to finance. One of the predictive models I designed was an Invoice Discrepancy Prediction model using a Multiclass Neural Network algorithm. This model predicts if an invoice will have a variance of some sort when checked against the purchase order, before the payments are to be processed.

Thanks to the model I designed, the productivity of processing invoices has increased by over 11%, because the team members only verify invoices that are discrepancy-free now.

• Split dataset
• variety of algorithms
• visualizing the data
• drag and drop capability 

are the features I appreciate most. 

The capability to model the data by finding empty cells and filling missing values by deriving the median and more, are great features that makes the job way easier.

I personally would prefer if data could be tunneled to my model through a SAP ERP system. It also needs features of Excel, such as Pivot Tables, integrated.

To create quick data analytic experiments, without incurring the time and cost of spinning up servers, setting up Hadoop, etc. 

Although MLS makes it very easy to deploy the resulting machine-learning models via REST API, I primarily use MLS as a means to quickly spin up experiments and create proof of concept models.

Not widely adopted at my old workplace, I only used this to create quick proofs of concept to try to convince management of the viability of a project.

MLS allows me to set up data experiments by running through various regression and other machine learning algorithms, with different data cleaning and treatment tools. All of this can be achieved via drag and drop, and a few clicks of the mouse.

The easy drag and drop can create simple data science experiments. Low barrier to entry allows large number of candidates get started.

The graphical nature of the output makes it very easy to create PowerPoint reports as well.

Enable creating ensemble models easier, adding more machine learning algorithms.

Out of about 150-plus MLS experiments I have done, maybe two or three bugged out. Interestingly enough, those are the ones I can’t delete out of the account.

Scalability, in terms of running experiments concurrently: Good. At max, I was able to run three different experiments concurrently.

Scalability in terms of deploying models: Unknown, I never deployed on Azure.  But I would guess REST API could probably easily handle a few K worth of hits per second, since that is how Microsoft is going to get paid.

Never used it.

The only other solution beyond this would be standard tools used by data scientists, like R, Python, etc. All of these would have a fairly high barrier to entry, requiring programming experience. The main selling point of MLS is the low barrier to entry, where even tech-savvy business people can use it.

Simple. Create MLS live account (preferably paid ones), open MLS, done.

Caveat: Different organizations have different attitudes towards cloud use, especially with sensitive data. At Bridgestone, the hardest part was getting corporate approval to allow me to upload heavily treated, sensitive data to a cloud platform.

To use MLS is fairly cheap. Even the paid account is something like $20/month,  unless you are provisioning large numbers of VMs for a Hadoop cluster.

The main MS makes money with this solution is forcing the user to deploy their model on REST API, and being charged each time the API is accessed. There are several pricing tiers for the API.

If you do not use the API, then value of MLS is to create rapid experiments ($20/month). The resulting model is not exportable to use, thus you’ll have to recreate the algorithms in either R or Python, which is what I did. MLS results gave me a direction to work with, the actual work is mostly done in R and Python outside of MLS.

R and Python.

Python + Pandas + scikit-learn: 

Pros: 

• scikit-learn offers better performance for extremely large data sets
• Large-data manipulation tools
• Fairly good set of ML algorithms

Cons:

• High barrier to entry, in terms of skill and knowledge
• Fairly labor intensive to create large number of experiments

R + caret:

Pros:

• Very good amount of ML algorithms (so many it may cause paralysis from too much choice, 200-plus algorithms)
• Good performance, unless the data set is extremely large

Cons:

• High barrier to entry
• Data manipulation is a pain, you probably want to use another tool to pre-treat the data before loading it into R dataframes

For data science professionals or programmers I would rate this solution a four out of 10. A major feature is missing: creating ensemble models. This can be achieved with the tool, but it's clumsy and slow.

For marketing or business professionals I would rate it an eight out of 10. It has a low barrier to entry, and can quickly create models that can be used for proof of concept and justify further investment in a full data science or Big Data project.

R and Python, in my mind, are still the way to go for a true data science/predictive analysis project. MLS's value is the ease of use and low barrier to entry. If one is not a programmer or statistician, MLS is a good way to get a project started, create a proof of concept.

I have used it to deploy predictive models in the healthcare sector.

It has helped in reducing the time involved for coding using R and/or Python. Also, web service is quite easy and convenient to use for clients. 

Its ability to publish a predictive model as a web based solution and integrate R and Python codes are amazing. It helps in building customized models, which are easy for clients to use.

It could use to add some more features in data transformation, time series and the text analytics section. Microsoft should also include more examples and tutorials for using this product.

• Feature-based selection
• Compute
• Data services.

Simplified development as scripts can be designed and implemented in real time.

I would like to see better prediction and analysis.

We have used it for a few months.

Good support is available when needed.

Stable at moment.

There are no scalability issues at the moment as data volume is still low.

Customer service is good.

Technical support is good.

We did not use a solution previous to this one.

It was complex to setup the workspace, but once it was done, we were good to go.

We did the implementation in-house.

The ROI was 36%.

The setup is a little complex, but it is worth it when it comes to security and efficiency.

We thought of doing this traditionally from scratch, but the Azure work space gives you the opportunity to utilize the environment and provide service in the shortest time possible.

For the best, reliable results, it is the best solution to have in mind. Try it out.

I use Azure Machine Learning Studio in my project to find solutions and build prototypes. It is mainly for fund management purposes and creating tools for specific cases.

The most valuable feature of Azure Machine Learning Studio for me is its convenience. I can quickly start using it without setting up the environment or buying a lot of devices.

Microsoft Azure Machine Learning Studio could improve in providing more efficient and cost-effective access to its tools for companies like mine.

I have been using Azure Machine Learning Studio for almost three years.

I would rate the stability of Azure Machine Learning Studio at around seven out of ten. Occasionally, there are minor hiccups, possibly related to bandwidth or server issues, but nothing significant.

The scalability of the solution is quite good and I would rate it as an eight out of ten. While it hasn't yet missed our requirements, we haven't pushed it to its limits. We don't deal with edge cases that demand extreme scalability.

Our clients typically include large multinational or state enterprises, as well as national companies in Indonesia.

Azure's setup feels friendlier and easier compared to AWS, making it simpler to understand and use. I would rate the easiness of the setup as an eight out of ten.

Deployment typically takes a few days to a few weeks to build prototypes and get familiar with available features. It is not too short to explore challenging cases, yet not too long to maintain efficiency.

I would rate the costliness of the solution as a nine out of ten.

I would recommend Azure Machine Learning Studio to others if they have enough resources to handle it. However, it is not a plug-and-play solution; there is a learning curve that needs to be addressed.

Overall, I would rate Azure Machine Learning Studio as an eight out of ten.