0.4.1 Laying the Groundwork: Installing Anaconda

There are many ways to install Python on our computer. One of the most popular ones is through the distributor/repository called . Anaconda (or simply Conda) is an open source distribution platform for Python and other languages created by Anaconda, Inc.—an American company founded in 2012. As of 2025, Anaconda claims it has about 50 million users worldwide. The reason for its popularity is that it serves as a package manager and it is free (though it also has a subscription-based premium version). Python is an object-oriented programming language. What makes it powerful is its rich and deep ecosystem of libraries, packages, and applications. Often, different packages and libraries are developed independently and may not necessarily play nice (compatibility-wise) with each other. These packages act as a patchwork in the fabric that is Python. Anaconda’s main service is to fit these patches together into “the Python quilt” and create environments where the libraries and native Python work well together.

We can download and install Anaconda from their free of charge. Anaconda provides download and installing documentation for the three major operating systems . Anaconda also provides a GUI called , which offers a simple and intuitive toolkit for installing and managing packages as shown in in .

A different way to download and install Anaconda without the Navigator GUI is to do so through a command line interface (or CLI), which is a program on our computer that processes text commands to do various tasks. Conda is a CLI program, which means it can only be used via the command line. On Windows computers, Anaconda recommends that we use the Anaconda Prompt CLI to work with Conda. This CLI is called Conda Powershell and it is included in the installation for free. On MacOS and Linux, users can use their built-in command line applications.

Anaconda contains a large suite of packages. There is a streamlined version called , which is a smaller version of Anaconda that includes only Conda, Python, the packages they depend on, and a small number of other packages. The smaller Miniconda allows for a streamlined installation, which we can then grow organically by installing packages as we need them. We discuss this practice below.

0.4.2 Getting Started: Step-by-step

Let’s get started by downloading and installing Miniconda.

• Turn our computer on and ensure we are connected to the internet.

• Open a web browser and navigate to the , or the .

shows what the installation website looks like in 2025. In the future, the layout of the website may look slightly different but the features and functionality should remain the same.

And scrolling down the same page, we can find the Miniconda installation button, which I highlighted in yellow. Click on it to install it on our computer. See .

shows that I downloaded Miniconda to my Windows 11 machine via Google Chrome. You can find it on your browser’s Downloads icon (highlighted in green). We can install it in our computer by clicking on the self-extracting link (highlighted in yellow).

Then we can follow the self installer, which will ask for a directory to save our Conda and Python environments to.

SILENT INSTALL

Another way to install Conda is through what is known as a “silent installation.” On a Windows 11 PC:

• Select the Start menu (the Windows icon) on the taskbar or press the Windows key.

• Type ‘cmd.’

• Select the Command Prompt from the list.

The command prompt will open to wherever the default path is located, which may look something like what is shown in .

Then type the following three commands:

These three lines instruct the computer to download the latest 64-bit Windows installer. We might want to rename it to a shorter file name or leave the three lines untouched. This performs a silent install, and then deletes the installer. It is called silent because it runs quietly in native MS-DOS, it does not open any screens, and all we can see as proof that it downloaded and installed is the following in the command prompt. See .

The command prompt now shows that 100% of the 88.9M Miniconda package was successfully installed on my Windows 11 machine.

There are additional ways to install Miniconda, such as from the Windows Powershell or directly from the Anaconda Navigator GUI. Instructions on these, as well as on how to install for Mac IOS and Linux are all available on the .

0.4.3 Working from the Miniconda Powershell

As we mentioned earlier, Python is a rich ecosystem of libraries and packages. These libraries are often developed independently, so they constitute a patchwork, and Conda is the quilt that gathers and patches up the various libraries. We will be doing analysis in an IDE. But before we can start working, we must set up an environment to do so. Miniconda is what we will use to set up the environment to work. Once we have installed Miniconda on our Windows 11 machine, we can simply go to the search box next to the Windows icon and type “miniconda”. See .

Clicking on the Miniconda Powershell is our first step into the world of Python. Clicking on it will open a window in our machine that will look reminiscent of an MS-DOS prompt. It will typically look like a black screen very similar to the Command Prompt screen shown in . Or it may look different if we have reformatted the look of the Conda Powershell in advance.

Since I have played with the look of my own Powershell in the past, my Conda Powershell does not look black. Regardless of how it looks, the prompt will automatically open wherever the Miniconda installation placed the base install on our computer. See .

At the Conda prompt, typing “conda - -version” will return the version of Conda installed in our computer. Our Miniconda installation would have already installed a few libraries along with the base installation. Typing “conda list” at the prompt will show the list of libraries originally installed with Miniconda. See .

We can always install more libraries onto our (base) installation as needed. However, I recommend keeping our base installation as light as possible. Why?

Let’s remember that Python libraries are often developed and updated independently by different parties—Who knows? You might create your own libraries yourself in the future? As a package manager, Conda patches up all these independent libraries and often creates dependencies that get these various libraries to play nice with each other. Since these Python libraries are independent, some libraries tend to be updated more regularly than others. This means that sometimes some libraries or packages get updated and no longer work nicely with the other libraries until other updates take place. So, it is a common occurrence that a project works one day and the next day, if a library gets updated, it stops working.

Here is where the utility of Conda really shines! Conda is not only a package manager, but also an environment manager. This means that we can set up an environment in Conda that is project specific. We can set up the libraries and packages we want under a given environment. And this environment acts as a freezer. If we do not update the libraries within that environment, the code is nearly guaranteed to continue to work even if versions subsequently change and libraries get updated.

Therefore, we can create different environments for different projects. We can create as many environments as we want and use them for different projects or for multipurpose. We can keep the libraries in an environment frozen or update them. We can share environments with friends. We can clone environments and we can delete environments when we no longer need them. Environments are our friends!

0.4.4 Python Environments with Conda

First, we want to find out whether we have Conda environments set up in our computer. At the Conda prompt, typing “conda env list” returns whatever environments we may have. If we just installed Conda, we likely have a single environment created with the base installation. We call this the (base) environment. See .

But again, a best practice is to leave (base) light and unencumbered with library/pckg bloat and, instead, to work in other environments. To do this, let us create the first environment. At the Conda prompt, type “conda create” followed by a space and a double hyphen and the word “name”. Whatever we write after the command “name” will be the name we are giving to the new environment. The name must be alphanumeric and it may not contain spaces. Let’s call our environment ‘Banking Intelligence 1’. This name will not work because of the spaces. An easy workaround is to connect words with underscores, hyphens, or periods.

To sum up, type “conda create - -name Banking_Intelligence_1. See .

This command gathers the necessary files and asks whether we want to proceed. Notice a warning that a newer version of Conda exists. (This will likely not appear on your end if you installed Conda recently.) But in any event, this will create the new environment without incident. Typing “y” for “yes” creates the package and returns us to the (base) prompt. See .

To verify that we did create our new environment, type again “conda env list” at the prompt and we can see two environments now. See .

Notice there is an asterisk next to (base), the asterisk denotes the environment that is currently active. Let us now activate our new environment. At the prompt, type “conda activate Banking_Intelligence_1” and then type again “conda env list” and we can see that our newly installed environment is now activated as we can see its name in the parenthesis at the prompt and we can see the asterisk next to the name in the environment list. See .

For practice, take some time now and create another environment called Banking Intelligence 2 and activate it. You should arrive at a new environment list that looks like .

Now let us say that in a few months we no longer need this newest environment and decide to delete it from your environments folder. This can be easily accomplished with the Conda command “remove.” However, we cannot delete the environment while the environment is active. We need to exit the environment first before removing it from the computer.

We can do it as before by activating another environment. For example, at the (Banking_Intelligence_2) prompt we could type “conda activate Banking_Intelligence_1” and we could then delete Banking_Intelligence_2 environment from there.

Alternatively, we could exit this environment by deactivating it, which would return us to the (base) environment by default. At the (Banking_Intelligence_2) prompt, we could type “conda deactivate Banking_Intelligence_2” and that will return us to base. See .

We can now safely delete this environment by typing “conda remove - -name Banking_Intelligence_2 - - all”. See .

Notice the environment is now gone from the list. As the environment was being deleted, Conda attempted to remove all packages, but it found none to remove. This is natural, since we had not installed any packages in that environment yet.

0.4.5 Python Libraries in Conda

Returning to our ‘Banking Intelligence 1’ environment, if we type “conda list” it returns an empty set since we have not yet installed any libraries in this new environment. See Figure .

Let’s now install the IDE we will use in this book. We will install the suite locally in our computer through the Conda Power shell. The Jupyter suite is a powerful IDE that enables interactive computing across many programming languages (the name is an amalgamation of the Julia, Python and R programming languages) in a web browser environment. Jupyter is an open-source, free software written in the open-standards paradigm, which means it is completely free of licensing and allows for free use, to be repurposed and freely distributed. This allows both the software and the code in this book to be available free of charge without any limitations or guarantees.

Jupyter suite provides a range of packages and libraries organized around two main IDEs: Jupyter Notebook and JupyterLab. Jupyter Notebook is the first IDE; therefore, it is older and possibly more popular. In recent years, JupyterLab is increasingly gaining popularity. All the codes in this book were written in JupyterLab. Those Python programmers (who often call themselves Pythonistas) who have been coding in Jupyter Notebook find transitioning to JupyterLab virtually seamless. If we are starting out fresh today, we need not begin from ‘Notebook’; we can begin to work and learn directly from ‘Lab’. One can do virtually the same things in both. The differences are mostly cosmetic, apart from the look of the ‘shell’ and a few additional toolbars and capabilities of the web browser one uses to work with the IDE.

One can install a package or a library in the Conda Powershell with the command conda install followed by the name of the library. Often Conda will install the desired package as well as all the required sub-packages, libraries, sub-libraries, and dependencies required for the package to run. At the (Banking_Intelligence_1) prompt of the Conda Power shell, typing “conda install jupyter” will gather all requisite components and ask to proceed([y],n)? See Figures and .

Typing ‘y’ will gather the packages and dependencies and install the Jupyter suite in the active environment. Once the suite and all its dependencies are installed, we can simply scroll down the shell to look for the libraries that were installed (the list is always organized alphabetically). See .

Alternatively, if we want to check whether a particular library or package is currently installed in the active environment, we can simply type the name of the package or library followed by two dashes and the word “version.” For example, at the prompt typing “jupyter - -version” returns the version of the libraries in the Jupyter suite. See .

We can easily see JupyterLab was also installed. Had it not been installed, we would need to type “conda install jupyterlab.” This would verify that the dependencies and libraries are all ready—if some libraries required some updating, Conda would prompt us whether we want to update them at this time or not—and subsequently it would install JupyterLab. Remember that this installation of the Jupyter suite has only been downloaded and installed in the currently active environment and not in (base) or in any other environment we might have created.

Two other libraries we will use often throughout the book are and . If we type “numpy - -version” or “pandas - -version,” we get an error if the library is not installed in the (Banking_Intelligence_1) environment. See .

At the prompt, we can easily install Numpy by typing “conda install numpy.” See .

Similarly, we can easily install Pandas by typing “conda install pandas.” Once we have completed this, we can verify these libraries are installed in the current environment by typing: “conda list numpy” and “conda list pandas.” See .

We will use Numpy and Pandas frequently throughout the book. We will also need other libraries and packages in this book and, later on, other projects you might undertake. Any additional library can be downloaded and installed in the same way as we described so far for Jupyter, Numpy and Pandas.

Sometimes, we may need a relatively large suite of libraries and packages, and (re-)installing them one-by-one may be time consuming. Let’s say that we have all the packages we need in our ‘Banking_Intelligence_1’ environment and we have been working in that environment for a while. Now, say, we want to create another environment—in our computer or in a different computer—that shares all the same libraries. Reinstalling each library and package one-by-one is always an option, but it can be time consuming. Also, the packages we might download today could be newly updated versions and, instead, we may want to keep the same versions of the libraries we had used before.

There are two ways to reinstall all the same versions of the libraries and packages onto a new environment, depending on whether we want the environment in the same machine or whether we want to create it in a different computer.

To easily replicate all the libraries we have in our environment onto a new environment in the same computer, we can simply clone the environment. Let’s say we want to create a new environment called Banking_Intelligence_5 and we want it to be identical to our Banking_Intelligence_1. This can be easily done at the Conda Powershell prompt by typing: “conda create - - Banking_Intelligence_5 - - clone Banking_Intelligence_1.”

0.4.6 (Re-)creating a Conda Environment via YAML

There is a somewhat more involved process to replicate an existing environment onto another computer. When working with Conda, a file can be used to list all the necessary packages for our project. This makes it easy to share our environment across different computers or with friends, which ensures all the necessary packages and correct versions are available in a new environment to run our code.

Starting from the environment we want to copy—say, ‘Banking_Intelligence_1’— we can export the current environment (with all libraries and packages in their current versions) by exporting it as .yml file in our computer. At the Conda Powershell, typing “conda env export > ‘name your new environment’.yml” will save a YAML file with the name we gave it locally in our computer—in the same directory where your environment is located. See .

This example creates a file I called “MyIB_env.” Typing “notepad myIB_env.yml” at the Conda prompt opens the YAML file, which contains a listing of all the libraries and packages installed in the “‘Banking_Intelligence_1’” environment. The “myIB_env.yml” YAML file is saved in the default folder. We can simply find it using the search box next to the start button on a Windows 11 machine. For example, the full path to my YAML file on my own computer was saved in “C:\Users\vicva\MyIB_env.yml.”

We can now create a new environment with all the libraries/packages reflected in the YAML file. Typically, we could do this by typing at the Conda Powershell prompt: “conda env create -f MyIB_env.yml.” However, notice that opening the “myIB_env.yml” YAML file in a notepad reveals the name of the original environment ‘Banking_Intelligence_1’. See .

This means that—while the name of the YAML file is “myIB_env.yml”—the name of the environment the YAML file calls is the original name “Banking_Intelligence_1.” So, if we attempt to create the environment from the YAML file, Conda will throw an error because the name of that environment (Banking_Intelligence_1) is already taken. But we can simply open the notepad, and right after the name field, type whatever new name we would like to give to the new environment and save the notepad. For example, let’s change the name in the name field from Banking_Intelligence_1 to Banking_Intelligence_7. See Figure .

We are now ready to create a new environment. At the Conda Powershell prompt type: “conda env create -f MyIB_env.yml.” See .

And voila! We can verify by typing “conda env list” that the new environment Banking_Intelligence_7, which is an exact replica of Banking_Intelligence_1, is now in our list of environments. See .

A YAML file with all the libraries required to run the codes for this book is available on the companion website .