machine learning - matplotlib error in matplotlib.backends import _macosx

When trying to visualize and plot data in Python you might work with Matplotlib. In case you are working on MacOS and you use a venv, in some cases you might run into the below error message:

RuntimeError: Python is not installed as a framework. The Mac OS X backend will not be able to function correctly if Python is not installed as a framework. See the Python documentation for more information on installing Python as a framework on Mac OS X. Please either reinstall Python as a framework, or try one of the other backends. If you are using (Ana)Conda please install python.app and replace the use of 'python' with 'pythonw'. See 'Working with Matplotlib on OSX' in the Matplotlib FAQ for more information.

The reason for this error is that Matplotlib is not able to find the correct backend. The most easy way to resolve this in a quick and dirty way is to add the following line to your code.

matplotlib.use('TkAgg')

This should remove (in most cases) the error and your code should be able to run correctly. 

Machine learning - Supervised machine learning and decision tree classifiers

When working with machine learning, and especially when you start learning machine learing one of the first things you will encounter is supervised machine learning and writing decision tree based classifiers. A supervised classifier which will leverage a decission tree to classify an object into a group will base itself on provided and already labled data.

The data will (in most cases) consist out of a number of features all describing labled objects in your training data. As an example, provided by Google, we could have a set of objects all havign the label apple or orange. In our case we have mapped apple to the numeric value 0 and orange to the numeric value 1.

The features in oour dataset are the weight of the object (it being an apple or an orange) and the type of skin. The skin of the object is either smooth (like that of an apple) or bumpy (like that of an orange). We have mapped bumpy to the value 0 and smooth to the value 1.

The below code showcases the implementation and also a prediction for a new object compared to the data we have in the learning data.

from sklearn import tree

features = [[140, 1], [130, 1], [150, 0], [170, 0]]
labels = [0, 0, 1, 1]

clf = tree.DecisionTreeClassifier()
clf = clf.fit(features, labels)

print clf.predict([[150, 0]])

As you can see in the above example we predict what the object with [150, 0] will be, will it be an apple or an orange. The above example is just a couple of lines and is already a first example of a simple machine learning implementation in Python. The reason it will only take this limited amount of lines is that we can leverage all the work already done by the developers of scikit-learn.

You can find the above code example and more examples on my Github project page.

[SOLVED] OSError: [Errno 2] "dot" not found in path.

When trying to visualize data using pydot in Python you might run into an error where it is stated that “dot” is not found in the path. This is after you already ensured you installed Pydot and ensured you did an import of Pydot in your python code. The main reason for this is that your python code is unable to find the dot executable. The dot executable comes from the graphviz project. This means that even though you did an install of the Pydot you are still missing a critical component.

If we look at the Pydot Pypi page you can see already a hint on this as it will tell you the following; Pydot is an interface to Graphviz and can parse and dump into the DOT language used by Grapgiz. Pydot is written in pure Python.

To resolve this is we can use yum to install Graphviz on Linux, in our case we use Oracle Linux.

yum -y install graphviz

this command will ensure that Graphviz is installed on your local Oracle Linux operating system, to check if the installation has been completed as expected you can use the below command to check the version;

[vagrant@localhost vagrant]$ dot -V
dot - graphviz version 2.30.1 (20180223.0356)

Now, if you run your python code and try something like pydot.graph_from_dot_data to work with dot data and visualize it at a later stage you will see you no longer have the issue you faced in the form of the OSError: [Errno 2] "dot" not found in path error faced before.

Oracle JET - data-bind as text or html

When using the KnockoutJS data-bind option while developing Oracle JET applications you will have the option to use the data-bind in combination with a text option or a html option. In most cases developers will use the text option to bind a javascript view/model variable in the html view code. In most cases variables will contain a value that needs to be displayed in the context of the html code, however, in some cases the variable can also contain html code itself. When a variable contains html markup code you will have to handle this differently when defining the data-bind in the view.

The below example screenshot displays the same variable used in a data-bind once with a text option and once with a html option. The first is using the text option and due to this the html code for the underline is not rendered. In the second line we use the html option and the underline markup code is rendered and the line is underlined in the view.

The below code example is showing the Oracle JET knockoutJS html view code, the code can also be seen as a gist on github.

The below code example is showing the Oracle JET knockoutJS html viewModel code, the code can also be seen as a gist on github

Oracle JET - Knockout data-bind

When using Oracle JET for building your application you automatically make use of Knockout.js as this is a vital part of how Oracle JET works. Part of Knockout is the data-bind option. Knockout’s declarative binding system provides a concise and powerful way to link data to the UI. It’s generally easy and obvious to bind to simple data properties or to use a single binding. Understanding data-bind is one of the basic parts you need to grasp to be able to develop an application with Oracle JET.

You can bind variables defined in the viewModel (javascript) and make sure they become part of the view (HTML) code.  In the below diagram you can see the binding between the Oracle JET View and the Oracle JET View/Model

Knockout data-bind example
As you can see in the below example screenshot we have displayed the value of someVariable0. The value of someVariable0 is set in the view/model

The below example code showcases the view/model, within IncidentsViewModel funtion you can see we assign the value to the variable which has been defined outside of the IncidentsViewModel function. You can view the example code as on GitHub via this Gist link.

To ensure we bind the IncidentsViewModel variable IncidentsViewModel defined in the view/model .js file we have to make sure we bind this in the html code of the view. The below example showcases how this binding is done as part of a HTML span using the data-bind option. You can view the example code as on GitHub via this Gist link.

In effect, the above example showcases the most basic form of a knockout based Oracle JET binding between the view and the view/model.

UX as part of your Enterprise Architecture

Digitalization within enterprises is still growing rapidly, enterprises are more and more adopting digitalization in every aspect of the daily processes and are moving to more intelligent and integrated systems. Even though a lot of work is being done in the backend systems and a lot of systems are developed and modernized to work in the new digital era a large part of the work has to do with UX User experience.

A large number of enterprises are still lacking in building a good and unified user experience for internal users. It has been thought for long that user experience was more applicable for the external systems such as websites, webshops and mobile applications. It is however evenly important to have a good and clear view on the internal user experience.

Internal user experience

Internal users, your employees, will use the systems developed on a daily basis. Ensuring the systems are simple to use, do what they promise and provide an intuitive experience will add to the productivity. Additionally, ensuring that systems are easy to work with and provide a good experience will ensure that your employees are more motivated and adoption of new systems will be higher

UX as an enterprise architecture component

In the past, it was common that every system within an enterprise would have a different experience. Menu structures, screen structures and the way a system behaved was different per application. As an employee normally interacts with multiple systems this can become overwhelming and complex. Additionally, it is relatively common that all internal enterprise user experiences are, to put it mildly, not that good. Most common, every system has a suboptimal interface and an interface design which is different from the rest.

An advised solution is to include standards for UX and interface design into the Enterprise Architecture repository and ensure, depending on your enterprise size, you have dedicated people to support developers and teams to include your enterprise UX blueprints within the internal applications.

When UX and interface design is a part of the enterprise architecture standards and you ensure all applications adhere to the standards the application landscape will start to become uniform. The additional advantage is that you can have a dedicated group of people who build UX components such as stylesheets, icons, fonts, javascripts and other components to be easily adopted and included by application development teams. At the same go, if you have dependency management done correctly, a change to a central UX component will automatically be adopted by all applications.

Having a Unified Enterprise UX is, from a user experience and adoption point of view one of the most important parts to ensure your digital strategy will succeed. 

Add UX consultants to your team

Not every developer is a UX consultant and not every UX consultant is a developer. Ensuring that your enterprise has a good UX team or a least a good UX consultant to support development teams can be of a large advantage. As per Paul Boag the eight biggest advantages of a UX consultant for your company are the following:

1. UX Consultants Help Better Understand Customers
2. UX Consultants Audit Websites
3. UX Consultants Prototype and Test Better Experiences
4. UX Consultants Will Establish Your Strategy
5. UX Consultants Help Implement Change
6. UX Consultants Educate and Inspire Colleagues
7. UX Consultants Create Design Systems
8. UX Consultants Will Help Incrementally Improve the Experience

Adopt a UX template

Building a UX strategy from scratch is complex and costly. A common seen approach for enterprises is that they adopt a template and strategy and use this as the foundation for their enterprise specific UX strategy.

As an example of enterprise UI and UX design, Oracle provides Alta UI which is a true enterprise grade user experience which you can adopt as part of your own enterprise UI and UX strategy. An example is shown below:

The benefit of adopting a UX strategy is that, when selected a mature implementation, a lot of the work is already done for you and as an enterprise you can benefit from a well thought through design. Style guides and other components are ready to be adopted and will not require a lot of customizations to be used within your enterprise so you can ensure all your applications have the same design and the same user experience. 

The above shown presentation from Andrejus Baranovskis showcases Oracle Alta UI Patterns for Enterprise Applications and Responsive UI Support

The adoption of chatbots in the enterprise market

We have seen the rise of chatbots in the past couple of years, more and more customer facing websites do implement a chatbot as part of the customer experience. Even though most people have had a negative experiences with chatbots the way they work is improving rapidly. Where chatbots used to be clumsy and not really good this is rapidly changing. The AI models behind chatbots is improving rapidly and they become more and more "human". As the maturity of chatbots is growing we see a growing adoption with chatbots by enterprises for both customer facing as well as internal facing communication.

As part of a Forbes article on the digital transformation trends in 2019 Chatbots have been placed on second place in the list.

1. Chatbots Good to Great: Hear me out on this one. I know we’ve all had extremely frustrating chatbot experiences as we round out 2018. But the good news is that huge steps continue to be made in the way of natural language processing and sentiment analytics—so many, in fact, that some believe NLP will shake up the entire service industry in ways we’ve never imagined. Think about all the services that could be provided without humans—fast food lines, loan processors, job recruiters! What’s more, NLP allows companies to gather insights and improve their service based on them. Some 40% of large businesses have or will adopt it by the end of 2019—which makes it one of our top 2019 digital transformation trends. Now, I know many are alarmed by where AI and Chatbots may impact the workforce, but I’m also bullish that companies are going to be upskilling their work forces rather than displacing them as machines may be good at delivering on clearcut requests but leave a lot to be desired when it comes to dealing with empathy and human emotion required to deliver great customer experiences."

Introducing a chatbot in the organisation
Enterprises in general are implementing chatbots for two main reasons; improving the efficiency to communicate with customers and improving internal processes. A commonly seen model is that enterprises take a two phase approach to introducing chatbots to the business.

Phase 1 - Internal use
In phase 1 chatbots are implemented and used to optimize internal processes. for example standard internal HR processes, supporting internal requisitions and internal IT support are commonly seen as first adopters of a internal enterprise chatbot.

Phase 2 - External use
In phase 2 chatbots are used externally facing as part of the enterprise website, shopping site or as part of enterprise mobile applications.

In general phase 1 and phase 2 overlap, while the go-live of phase 1 is in effect phase 2 is already being prepared for external use. By creating the correct overlap the momentum of the chatbot team is maintained and the lessons learned from phase 1 are included in phase 2. It is important from both a team velocity as well as an adoption point of view to ensure you keep the momentum and ensure an overlap or a minimal gap between phase 1 and phase 2.

It is not done in a day
Contradicting the popular believe that building and implementing a chatbot is an easy task one will have to prepare for a "real project". Even though the use of a cloud platform and chatbot framework can speed the technical implementation up extremely a healthy part of the work is in ensuring your chatbot has the correct vocabulary and ensuring your conversation design is properly done.

Two aspects are important when developing your chatbot project planning. The first is to ensure you enough space for conversation design and ensuring the right vocabulary. Conversation design will go into design of how a flow of a conversation between your bot and a human will go. Even though this might sound straightforward initially it might be a very good practice to ensure you have an experienced conversation design expert on your team.

The other important part is to include a marity model for your chatbot in your project planning and strategy. The moment you want to launch internal and the moment you want to launch externally might be on a different point in the maturity model. An example of a chatbot maturity model, developed by Leon Smiers at Capgemini, can be seen below.

Use a chatbot framework
Building a chatbot from the ground up, building all the AI and all the other parts needed to make a good chatbot is an amazing project. However, such a project is only interesting from a technical understanding and research point of view and not so much from a business point of view. As a developer who just wants to build and include a chatbot interaction it is a better solution to leverage an existing platform. As an Example, Oracle provides a intelligent chatbot platform.

The below developer conference video showcases how to build a chatbot.

You can find more information and developer code examples via this link to get started quickly with your first intelligent chatbot to include in your enterprise landscape.