Microsoft Azure Machine Learning Studio Reviews

Our use cases  involve customer segmentation for targeted marketing, where I use machine learning to identify potential customers interested in a new product. Another is a recommendation system on our company website, where I use machine learning to suggest additional products to customers based on their browsing or purchase history. Lastly, there is pricing estimation, where I use machine learning to predict the price of an item or article.

The features of Azure Machine Learning Studio that I find most valuable depend on the type of model I'm working with. For integration, knowing halfway indicators is crucial to assess model performance. For classification models, the confusion matrix is important for evaluation, while for regression models, statistical tests like the provision statistics are valuable.

One area where Azure Machine Learning Studio could improve is its user interface structure. Simplifying the initial information presented upon first use could make it more accessible, especially for users with limited technical skills. Providing only essential information upfront would enhance the user experience and reduce complexity.

I have been working with Microsoft Azure Machine Learning Studio for three years.

I would rate the stability of the solution at a six out of ten. Improving stability involves finding people with the right skills to handle problems that arise. While stability depends on how well the solution is installed, ongoing efforts are needed to address issues and refine the system. We are working step by step to identify and solve problems, but there's room to find more comprehensive solutions as they come up.

I would rate the scalability of Azure Machine Learning Studio at about a seven out of ten. While it offers high scalability, it can be challenging for less technical users and may encounter issues with defects and industrial licensing, particularly in logistics projects.

At our company, we use Azure Machine Learning Studio daily.

Before Microsoft Azure Machine Learning Studio, we used on-premises solutions. We made the switch to Azure Machine Learning and the cloud to modernize our projects and leverage the benefits of cloud computing.

I use Azure Machine Learning Studio for predictive modeling in my project. I follow a workflow that involves selecting data, preprocessing it, training models, and deploying them. The Studio's tools cover all these steps, making it convenient for me to build and deploy predictive models.

In a specific scenario, I used Azure Machine Learning Studio for data preprocessing by creating new variables. This involved tasks like transforming variable types or combining multiple variables to create new ones. Additionally, I employed cross-validation techniques, such as k-fold validation, to assess model performance and select appropriate metrics for evaluation.

The most important aspect of my machine learning projects is the quality of the data. It is crucial to determine whether the data can provide meaningful information relevant to the project's use case, regardless of the specific tools or features used.

The drag-and-drop interface of Azure Machine Learning Studio has greatly improved my workflow. It is easy to use and increases productivity by allowing quick experimentation and visualization of data pipelines. This feature enables me to iterate rapidly and efficiently, especially for small projects or presentations.

I would rate the performance of the solution at an eight out of ten for my team. However, our data volume is not the largest. While I believe our performance is strong, other companies might rate it lower due to different circumstances.

My advice for someone considering installing Azure Machine Learning Studio is that it is user-friendly, especially for technical users. You can easily upload data and analyze it with the examples provided. The drag-and-drop interface makes it intuitive, and upgrading to this tool for data analysis is a good idea.

Overall, I would rate Azure Machine Learning Studio as a nine out of ten.

My company uses Microsoft Azure Machine Learning Studio to help our company's customers view AI solutions.

My company's clients' use cases will be that they use the solution to feed information to the system about their customers who purchase from them. The solution also helps one to combine products to engage in cross-selling and upselling activities while keeping track of customer lifetime value. The solution also helps its users with the pricing simulation part to figure out what prices are good for the business and maximize the closing of the sale.

The most valuable feature of the solution is the availability of ChatGPT in the solution.

Improvement will be possible with more machine learning functionalities in Microsoft Azure Machine Learning Studio since, at times, the current accuracy of the solution is not good enough. It would be good if Microsoft Azure Machine Learning Studio could have a generative AI tool similar to ChatGPT.

I have been using Microsoft Azure Machine Learning Studio for three years. My company functions as a reseller and a partner of Microsoft.

Stability-wise, I rate the solution a seven out of ten. With ML, you may face some data-related issues, especially considering that when dealing with customers at times, the data that comes in might not be clean. Stability-wise, you may face certain problems when you fail to refresh the data in the solution.

Scalability-wise, I rate the solution a seven out of ten. My company still has to do some of our own optimizations to the data part of the solution until and unless we subscribe to some third-party data lake services, which is a better option but comes at a higher cost.

My company's client's organization has around 10 to 50 users of the solution.

My company caters to the requirements of medium and enterprise-sized companies.

I rate the technical support a six out of ten.

Neutral

I rate the initial setup phase of the solution a six on a scale of one to ten, where one is difficult, and ten is easy. The initial setup phase of the solution was a bit complex. The setup phase is a bit difficult if you want to view Microsoft Azure Machine Learning Studio as an application.

The solution is deployed 50 percent on the cloud and 50 percent on-premises.

Considering the fact that my company currently builds some standard solutions, Microsoft Azure Machine Learning Studio's deployment takes us around two to three months.

I rate the solution's pricing a four on a scale of one to ten, where one is cheap, and ten is expensive.

There are some additional payments to be made apart from the licensing fees of the solution since buying Microsoft Azure Machine Learning Studio alone won't make it a complete solution. You will need the database and data lake services.

Microsoft Azure Machine Learning Studio does not allow users to have a PnP option, like an ERP or a CRM system, where everything works if you include the data with the system. Sometimes, it is difficult to generate good patterns using the solution. You need to have good experience with the solution to move around with the data from the beginning before coming up with different strategies to end different problems. In general, the product is not a straightforward solution.

There is a need for Microsoft Azure Machine Learning Studio's users to put in some programming efforts to make the solution work accurately under different scenarios.

I rate the overall solution a six out of ten.

In terms of what I found most valuable in Microsoft Azure Machine Learning Studio, I especially love the designer because you can just drag and drop items there and apply the logic that's already available with the designer. I love that I can use the libraries in Microsoft Azure Machine Learning Studio, so I don't have to search for the algorithms and all the relevant libraries because I can see them directly on the designer just by dragging and dropping. Though there's a bit of work during data cleansing, that's normal and can't be avoided. At least it's easy to find the relevant algorithm, apply that algorithm to the data, then get the desired output through Microsoft Azure Machine Learning Studio.

I also like the API feature of the solution which is readily available for me to expose the output to any consuming application, so that takes out a lot of headache. Otherwise, I have to have a developer who knows the API, and I have to have an API app, so all that is completely taken care of by the Microsoft Azure Machine Learning Studio designer. With the solution, I can concentrate on how to improve the data quality to get quality recommendations, so  this lets me concentrate on my job rather than focusing on the regular development of APIs or the pipelines, in particular, the data pipelines pulling the data from other sources. All the data is taken care of and you can also concentrate on other required auxiliary activities rather than just concentrating on machine learning.

As for the areas for improvement in Microsoft Azure Machine Learning Studio, I've provided feedback to Microsoft. My company is a Gold Partner of Microsoft, so I provided my feedback in another forum. Right now, it is the number of algorithms available in the designer that has to be improved, though I'm sure Microsoft does it regularly.

When you take a use case approach, Microsoft has done that in a lot of places, but not on the Microsoft Azure Machine Learning Studio designer. When I say use case basis, I meant recommending a product or recommending similar products, so if Microsoft can list out use cases and give me a template, it will save me a lot of time and a lot of work because I don't have to scratch my head on which algorithm is better, and I can go with what's recommended by Microsoft.

I'm sure that isn't a big task for the Microsoft team who must have seen thousands of use cases already, so out of that experience if the team can come up with a standard template, I'm sure it'll help a lot of organizations cut down on the development time, as well as going with the best industry-standard algorithms rather than experimenting with mine.

What I'd like to see in the next version of Microsoft Azure Machine Learning Studio, apart from the use case template, is the improvement of the availability of libraries. Microsoft should also upgrade the Python versions because the old version of Python is still supported and it takes time for Microsoft to upgrade the support for Python. The pace of upgrading Python versions of Microsoft Azure Machine Learning Studio and making those libraries available should be sped up or increased.

I've been working with Microsoft Azure Machine Learning Studio for nearly two years now.

Microsoft Azure Machine Learning Studio is a stable solution. My company is already using it in production. At least customers use the recommendations from Microsoft Azure Machine Learning Studio in production, so the solution is quite stable, at least in cases developed by my company.

Microsoft Azure Machine Learning Studio is a solution that's easy to scale. It's pretty easy because it is hosted on Kubernetes, and there is an option in the portal where I can simply move my plan from standard to enterprise. The solution also has an automatic scaling option available because it is on Kubernetes, so it can scale automatically. I'm seeing that it's quite scalable. This has nothing to do with availability because it just runs in the background, and it is not customer-facing, but the output is customer-facing, so availability is a different case, but in terms of scalability, Microsoft Azure Machine Learning Studio is scalable.

The technical support team for Microsoft Azure Machine Learning Studio was pretty good, though I had to tailor the answers to my requirement, but would rate support a four out of five. Most of the questions my company had, more or less, the support team already experienced, so the team had answers readily available which means there wasn't a need to do a lot of R&D, so getting answers from technical support didn't take a lot of time.

In terms of setting up Microsoft Azure Machine Learning Studio, initially, when my company started, the documentation wasn't so good, but now it has improved. Provisioning the solution only takes a few clicks, so it's no big deal, but setting up the pipelines because no enterprise will have a single environment, you'll have to create multiple pre-production and end production environments, so moving my latest changes to the next environment was a bit of a challenge.

Many terminologies are now in the market such as DevSecOps, and MLOps, so that MLOps documentation was available initially, but it wasn't very explanatory, but now, there's a lot of improvement in the MLOps documentation and that will help me move and propagate my changes from one environment to another.

Microsoft has made improvements into the tutorials, especially on MLOps. Finding MLOps experts in the market was also very tough initially, so my company was trying to learn on the job and do it, so it took some thinking and time, but it's still good because you can learn on the job and do it, but you won't always have the luxury of time to learn it.

In terms of pricing, for any cloud solution, you should know the tricks of the trade and how to use it, otherwise, you'll end up paying a lot of money irrespective of the cloud provider, so at least for Microsoft Azure Machine Learning Studio pricing versus AWS, I would rate it three out of five, with one being the most expensive, and five being the cheapest. It could be cheaper, but you also have to be careful when choosing the plans, for example, consider the architecture and a lot of other factors before choosing your plan, if you don't want to end up paying more. If your cloud provider has an optimizer that seems to be available in every provider, that would keep alerting you in terms of resources not being used as much, then that would help you with budgeting.

We evaluated quite a lot of options. We compared Microsoft Azure Machine Learning Studio against Google Cloud and AWS solutions, and there were several others available in the market. I'm trying to recollect the names which we compared the solution with. We did the benchmarking, but we went with Microsoft Azure Machine Learning Studio because our clients and their data were on Azure, though that doesn't necessarily make you go with the solution. After all, you can pull the data from any other cloud as well. For our use case, however, we found many of the things were readily available and the learning curve for Microsoft Azure Machine Learning Studio compared to others was better and easier. We didn't have to search for experts in the market to hire them because we could have our in-house team learn and deliver the solution on the job.

Microsoft Azure Machine Learning Studio is a cloud-native solution. It's completely cloud-based.

My company has eight users of Microsoft Azure Machine Learning Studio.

My rating for Microsoft Azure Machine Learning Studio is seven out of ten.

Developing and operationally implementing a powerful lead scoring model for a major Multiufamily developer and operator of apartment properties throughout major western states. The work included 3 years of data across over 60 properties with more than 500,000 leads and 3 million transactions.

Increased sales force productivity by permitting them to prioritize activity during peak leasing periods on those leads most likely to close

The ability to do the templating and be able to transfer it so that I can easily do multiple types of models and data mining is a valuable aspect of this solution. You only have to set up the flows, the templates, and the data once and then you can make modifications and test different segmentations throughout.

We were working across a number of internal departments as well as some outside departments and this solution made it extremely easy to communicate across functional area because it was all in flow chart and data form so that if somebody had an issue, like changing the data set or something like that, they could point right to it and we could get that handled and incorporated into the model. It's extremely efficient on the computer. We had to do a number of resets on the data in the model and to be able to turn things around and validate the model and the new set in two hours, was just incredible for me.

It was very robust. The ability to move the objects around so easily and then communicate is really its power. Then to be able to show it to the sales and senior management, in terms of what was employed and made it very easy to get my job done.

In terms of improvement, I'd like to have more ability to understand the detailed impact of the variables on the model and their interactions. Their algorithms are very powerful and they explain overall the net contribution of each of the variables to the solution. In terms of being able to say to people "If you did this, you'll get this much more improvement" Azure (at least my understanding of it) doesn't provide readily accessible tools to assess from a management perspective the impact of their changing a sinimized, the better.gle value - for instance in closing a lead, decreasing response time by 10%.

I recognize that the multivariate algorithms used from decision trees to neural nets do not readily provide the coefficients for each variable ala the older regression modeling approaches. My experience over my 50 years of developing and implementing predictive models has been that more than half the value of modeling lies in improving management's understanding of the process being modeled, often leading to major organization and operational structure changes. More ability to understand the variables impacting the end result being optimized would be very useful. 

I have worked extensively with this solution for the last three years. 

I haven't had any problems with stability. 

I didn't have any issues with the scale. we rapidly went from test to full implementation across all datasets.

I never had to use technical support.

I have used SPSS modeler (part of WATSON really) but because client was a Microsoft shop, I switched to Azure.

I found the setup to be very easy. I've been doing this type of work for 50 years so the modern terminology isn't always the same as what I grew up with. It took me a while to understand that, but the setups were very easy. As with anything, the hardest part is always getting the data together, but the outside consultants had built up a very, very good data warehouse. The ability to manipulate the data and create variables was very nice.

THIS IS THE ONLY MODELING APPROACH THAT EVER WORKED THE VERY TIME I RAN IT!!

Because client isa Microsoft shop, everything was Microsoft in terms of having solutions like Power BI and stuff like that. Azure is very useful and very inexpensive.

The major advice I give is that clients must get the user,somebody who understands the business issues, to be deeply involved with it and the data transformation. Most people don't. And that's true for data science applications. We don't just follow the data in a big pile and remodel, we advance the process that we're modeling. Consider what transformations of the data you need to make it workable and usable.

Remember, over half the initial value of modeling is the strategic understanding provided re the importance of different variables to the model and hence the organizaion's performance. Very often the modeling identifies opportunities for changing structures, decision rules, etc. even prior to the model's actual implementation technically.

I would rate it a nine out of ten.

We used this product as implementers. For example, a client wanted to use the Azure stack, specifically Azure Machine Learning Studio. Microsoft was a consultant on the project, and we were the implementation partner.

There are actually two tools. One is Azure Machine Learning Designer (which used to be called Azure Machine Learning Studio), and the other is Azure Machine Learning. Designer is a drag-and-drop interface, primarily for those without extensive coding expertise.

Azure Machine Learning has become the de facto product, and it allows you to write code and provides numerous components for building machine learning models.

We build all our machine-learning models using ML Studio. That's the primary purpose for us. We build standard models that use data and various algorithms.

My expertise was focused on the analytics aspect. We worked in a large firm with different teams responsible for MLOps and integrating outputs with other systems. 

My responsibility was to get data, build models, and provide the output. How the output was rendered and the process of doing so was handled by other teams. 

Additionally, there's a separate team that handles the productionalization of machine learning models. It's a diverse team structure, so the scope of the totality of the integration process varied. Essentially, they take the model I build, use Azure DevOps for versioning, and then deploy it to production.  

Everything in Azure is very intuitive. As a Microsoft product, it's designed that way. You can easily search for things, and many features are built-in. Many things can be easily done via drag-and-drop.

We often use custom code that we can template or create as boilerplate code for different teams. It sounds difficult, but Azure makes it very easy. 

The learning curve is very low. Operationalizing the model is also very easy within the Azure ecosystem.

The drag-and-drop interface specifically improves our workflow a lot. Designer simplifies the process when we want to create a quick model. It makes things very, very easy. It's a great starting point.

When we want to create models, model management and deployment are crucial. We had boilerplate code ready, and Azure made it significantly easier to deploy using its services. That was the easy part.

In terms of data capabilities, if we compare it to Google Cloud's BigQuery, we find a difference. When fetching data from web traffic, Google can do a lot of processing with small queries or functions. Azure didn't have that same inbuilt feature for website traffic or analytics, unlike Google DB and BigQuery.

I started using it around the year 2020. We used it up to 2023, so roughly two years.

It is a stable product. I would rate the stability a nine out of ten. 

It is a scalable solution because it works well for large databases.

I would rate the scalability a nine out of ten. 

The customer service and support are good. 

Positive

We've used Google Vertex AI and AWS SageMaker.

I was not involved in the deployment process. But there is maintenance. It was quite a headache. 

Maintenance does require attention. With any cloud implementation, cost optimization is a major factor. Our team had discussions about it. 

I would rate the pricing an eight out of ten, with ten being very expensive. Not very expensive, not very cheap.

It was on a yearly basis, and there were also usage-based costs.

If it's my recommendation, it's a very competent product. It has all the necessary features for data engineering, data science, and model management. 

It's a complete suite of products that can address your end-to-end data needs.

Overall, I would rate the solution a nine out of ten. 

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.

We use Microsoft Azure Machine Learning Studio to generate predictive sales analytics and determine customer behavior.

Through the solution's customer data analysis, we conduct customer data experiments, test hypotheses, and develop sales strategies.

The solution is integrated with our Microsoft Azure tenant, and we don't have to go anywhere else outside the tenant. The solution's data pipelines are easier to configure, and the solution provides a range of tools and libraries we can access.

It would be great if the solution integrated Microsoft Copilot, its AI helper.

I have been using Microsoft Azure Machine Learning Studio for one year.

The solution's stability depends on the fragility of libraries and the availability of services. Sometimes, the demand is very high in the public cloud, and performance and availability issues have occurred.

I rate the solution a six out of ten for stability.

Microsoft Azure Machine Learning Studio is a very scalable solution. Three people are using the solution in our organization.

I rate the solution an eight out of ten for scalability.

I rate the solution a seven out of ten for the ease of its initial setup.

The solution’s deployment takes one hour.

There is a lack of certainty with the solution's pricing. The risk is the pricing is high without you necessarily knowing. The workload drives the solution's pricing. If you give it a lot to do, it will cost a lot of money. It's about committing to how much you want to pay for. You don't necessarily know what you'll get for the price level that you agree.

On a scale from one to ten, where one is cheap and ten is expensive, I rate the solution's pricing a seven out of ten.

Before choosing the solution, we evaluated Databricks. We chose Microsoft Azure Machine Learning Studio to get as close to the Microsoft pattern as possible. We have a Microsoft first policy, and therefore, unless there's a reason not to use Microsoft, we choose Microsoft.

I would recommend Microsoft Azure Machine Learning Studio to other users. I would also ask users to compare the solution with Microsoft Fabric, which is a collection of components to put a workflow together end to end.

Overall, I rate Microsoft Azure Machine Learning Studio a seven out of ten.

Microsoft Azure Machine Learning Studio can be used for developing models, such as predicting energy usage, as I did for my bachelor's project, where I predicted future energy usage for a city in Norway. The solution can also be used for classification tasks, such as identifying objects in images.

In terms of features, I personally find Azure to be clearer and better than Google because it provides better quality and clarity regarding what needs to be done.

The icons in the solution could be improved to include examples of how to use each container, as sometimes it's unclear which container to choose. It would be helpful to provide examples to understand better which virtual machine or how many courses to use. Overall, the icons in the solution could be improved to provide better guidance to users.

Additionally, the setup process for the solution could be made easier.

I have been using Microsoft Azure Machine Learning Studio for half a year. I am a student and user of the solution.

I think Microsoft Azure Machine Learning Studio is more stable than Google.

In terms of scalability, I believe that the solution is good. Although I have only used it for two projects, I think that it provides a good level of scalability. However, as I have only used it within my organization, I may not have experienced all of the possibilities that the solution offers.

Regarding the technical support for the solution, I find the documentation provided comprehensive and helpful. It is often the case that everything one needs is already in the documentation, so I haven't had to use the support much. Even when I have reached out for support, I have always received a prompt response.

The initial setup for me was initially quite complex, but after completing a course related to Microsoft Azure Machine Learning Studio, it became less complex. However, one needs to have a good understanding of the required parameters and what the model needs to do in order to achieve good performance. So sometimes, it's not that simple. The deployment process took me a couple of hours to complete. I was able to do it quickly because I was using Azure Machine Learning Designer and Python SDK while also learning automation. The setup process for AltaML was easy and could be completed in hours. With Python SDK, the setup process was quite long because of the code that needed to be written, so one needs to know what to write.

I used the free student license for a few months to operate the solution, but I'll have to pay for it if I want to do more now.

Before choosing Microsoft Azure Machine Learning Studio, I only evaluated Google Cloudpath.

If you plan to use this solution, I suggest you not be intimidated by its complexity at first. You will gain more clarity regarding the solution over time with perseverance and practice. Overall, I rate the solution an eight out of ten.