<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Jennifer Arty on Medium]]></title>
        <description><![CDATA[Stories by Jennifer Arty on Medium]]></description>
        <link>https://medium.com/@jennifer.arty?source=rss-2dba6326e7b5------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*dmbNkD5D-u45r44go_cf0g.png</url>
            <title>Stories by Jennifer Arty on Medium</title>
            <link>https://medium.com/@jennifer.arty?source=rss-2dba6326e7b5------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 08:31:24 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@jennifer.arty/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Creating Word Clouds]]></title>
            <link>https://medium.com/@jennifer.arty/creating-word-clouds-b3aa2d0f5e8e?source=rss-2dba6326e7b5------2</link>
            <guid isPermaLink="false">https://medium.com/p/b3aa2d0f5e8e</guid>
            <dc:creator><![CDATA[Jennifer Arty]]></dc:creator>
            <pubDate>Thu, 08 Nov 2018 18:39:47 GMT</pubDate>
            <atom:updated>2018-11-08T18:39:47.489Z</atom:updated>
            <content:encoded><![CDATA[<p>Creating word clouds can be a fun, exciting and great way of visualizing text. I recently discovered how to create word clouds while thinking of ways to add a fun visual about my last Data Science project. Using word clouds was a quick and easy way of showing us the most used words with visualization.</p><p>There are a couple of things that you need to get the word cloud to generate. First, you need to import the following libraries into Python to read the PNG file that will be used to create your word cloud. Second, you use Numpy to open and read the PNG file to generate your word cloud.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ctRmPJWxzu5P0yvUAjspYg.png" /><figcaption>importing libraries into python</figcaption></figure><p>Next, you need to create a list of texts for the image, and import stopwords from the word cloud library. Wordcloud comes with generic stop words already incorporate into their library, but it is also possible to add more stop words using stopwords.udpate([‘list of words’].</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*PzIJutQ-o1vPaF6mtbFRng.png" /><figcaption>generating word list and stopwords</figcaption></figure><p>Now its time to import our image and start to format it! Import the image using np.array(Image.open). It will generate a np.array with 0,1,2 (black ) based on the type of PNG file. Then convert these outputs to 225, using format_wine function since this will give a mask of white instead of the black that was originally showing.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fCaLQfPxSyGbtoNn9WH5ig.png" /></figure><p>After building the mask, you can now generate the image using the built-in WorldCloud function. It comes with many different parameters to customize your image. This is where we add our mask, edit the background color, font, stopwords, etc.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NaQZ_BIHRWaQ2DlLiX69_g.png" /><figcaption>WordCloud Parameters</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*xp3VPSUwLtBHdZjhJy4YDg.png" /><figcaption>Generating word cloud, storing it and showing the image</figcaption></figure><p>Tada! Now we generated a word cloud using a simple PNG image of a wine bottle and glass, to using keywords that are used in wine reviews.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*zyoXBKph_bmrPH-r5Ka7jA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/802/1*tbL9xABBzSSn1RnhztZgmQ.png" /></figure><p>One thing I found interesting was using interpolation in plt.imshow function. Based on Image size, you can adjust the way the displayed words by changing the interpolation. Below is demonstrating the interpolation based off of ‘bilinear,’ None and ‘Sinc.’ We don’t see too many differences between the three options besides the slight change in colors and boldness of the words.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*c2Fs1xuNLFWig0eDMfTGVg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1020/1*4amQOB4mMSxokCYwOHP5jg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/938/1*FnHyVFLzT33u0tJ-gOTFrQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*E_RmaUVutY1xxulSPoeE5g.png" /></figure><p>You can now use word clouds one almost any PNG file, which gives an unlimited amount of creativity to projects. Below I used the same wine database reviews to generate the word cloud based off the NYT logo. Some changes I made:</p><ul><li>Incorporated a margin to fit more words into the logo</li><li>Added a border using count_color = “black”</li><li>Changed the default word colors to black using the color function</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ugcnbQZIq-_B8KYRl3RZaA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/718/1*gEMI1iu6BZcM8gO_tmjOag.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b3aa2d0f5e8e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Exploring Convolutional Neural Networks]]></title>
            <link>https://medium.com/@jennifer.arty/exploring-convolutional-neural-networks-39cd4f23080?source=rss-2dba6326e7b5------2</link>
            <guid isPermaLink="false">https://medium.com/p/39cd4f23080</guid>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Jennifer Arty]]></dc:creator>
            <pubDate>Mon, 08 Oct 2018 14:03:14 GMT</pubDate>
            <atom:updated>2018-10-08T14:03:14.289Z</atom:updated>
            <content:encoded><![CDATA[<p>I recently learned about Convolutional Neural Networks (CNN) during my last module at the Flatiron School Data Science program. I worked on a project to use CNN to recognize different types of food. CNN’s are typically used for image processing due adding the convulsions layers, where the respective neurons are only connected to pixels in the designated field. This lets the network focus on smaller features in image, allowing for better image recognition.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*RtZCRwjVNV8cJMoc.png" /></figure><p>For this project I used the Food 101 dataset, which has 101 different food categories each containing 1,000 pictures. Since it was my first time using CNN to identify pictures, I started off small seeing if the network could classify if a picture was sushi or of nachos.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/384/1*9KXBcX8iGdoxjl1vWYm_VA.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*Lsb_l92r97t97gQJ-ksiWw.jpeg" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/382/1*k6e3A2ewMZjWvuOqgCvS0w.jpeg" /><figcaption>Food 101 dataset of sushi and nachos</figcaption></figure><h3>Train/Test and Augmentation</h3><p>First step was to set up the training and testing images. The training images contain 600 pictures of each sushi and nachos, while the test images contained 200 pictures of each sushi and nachos. To make sure that the network was being trained on as many different scenarios of images as possible, I used image generator to reshape, flip, scale etc. each image in the set. This allows the CNN to see the same picture in different ways in order to be able accurately classify the images.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/544/1*MyxFBvOwAPRcZf6PYLLCew.png" /><figcaption>Example of augmented images</figcaption></figure><h3>Setting up CNN</h3><p>I used Kera’s Sequential model with 8 different layers, alternating from Conv2D (Convolutional layer of 2-D inputs)and MaxPooling. To minimize any overfitting int he training model, I implement drop out regularization at 50% in the last later. Sigmoid activation function was used since I have a binary model.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1_Dl7K0sbZZOHEnIKjvzCw.png" /><figcaption>Convolutional Neural Network Layers</figcaption></figure><h3>Train/Test and next steps.</h3><p>Time for training and testing! The CNN was ran for 30 epochs resulting in a training accuracy of 97.5% and a test accuracy of 81.2%.The model began to overfit around 10 epochs. Applying regularization method of stop early at 10 epochs the model finished with a training accuracy of 89% and test accuracy of 80%.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/438/1*dguV338bux05KTFZ3byMmg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/622/1*1goVi9NLrpK7RI-iUUzl8w.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/454/1*OmxFBMxHsASjTsso7K1WLA.png" /><figcaption>Training/Test for 30 epochs</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/352/1*jmK3750z7NkgHQqgq9r8cA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/606/1*eeY1V2C8LarugDYxeskpeA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/364/1*pMaxuQWzAuX6Zp92jw_0Yg.png" /><figcaption>Training/ Test results for 10 Epochs</figcaption></figure><p>My model performed well, but there is still room for improvement by adding more layers and implementing transfer learning. I also tried to add additional categories, but did not see as great of a results. Stay tuned for my next post about how my experience with adding more food categories and implementing transfer learning!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=39cd4f23080" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Regularization methods to prevent overfitting in Neural Networks]]></title>
            <link>https://medium.com/@jennifer.arty/regularization-methods-to-prevent-overfitting-in-neural-networks-1a79b5e3081f?source=rss-2dba6326e7b5------2</link>
            <guid isPermaLink="false">https://medium.com/p/1a79b5e3081f</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[neural-networks]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Jennifer Arty]]></dc:creator>
            <pubDate>Tue, 25 Sep 2018 11:37:57 GMT</pubDate>
            <atom:updated>2018-10-08T14:51:49.129Z</atom:updated>
            <content:encoded><![CDATA[<p>Overfitting is a common problem that all Machine Learning Algorithms run into. This occurs when the model is fit too well to the training set but does not perform as well on the test set. Models can also run in to the issue of under fitting, when we choose too simple of a model that it does not perform well on both the training and test set. For Deep Learning, the issue of overfitting since the model is able to adapt freely to large amount of complex data. The most common ways to reduce overfitting is through Regularization methods of Early Stopping, L1, L2 and DropOut.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/555/1*tBErXYVvTw2jSUYK7thU2A.png" /></figure><p><strong>Early Stopping: </strong>stopping the model when thee validation model reaches a plateau.</p><p><strong>L1: </strong>shrinking less important coefficients to zero, removing these features entirely from the model. (Lasso)</p><p><strong>L2: </strong>reducing coefficients to close as zero, reducing the importance of less important features.(Ridge)</p><p><strong>Drop Out : </strong>every neuron as the probability (defined by the user) of not being used during a round of training, but will be implemented in the next round. This allows the neurons that are training to increase their learning capabilities.</p><p>To demonstrate these methods, I am using the Wine Review data set from Kaggle, to test out the different data sets. My goal is to find out the category of wine based off of the wine review. I started by cleaning the data and converting the wines into 19 different categories ranging from Dry Red — Champagne. With taking a random sample of 10,000 wine reviews in order to find the category of the wine.</p><h3><strong>Early Stopping</strong></h3><p>The first regularization method I am going to try out is early stopping, which means to stop the model as soon as the validation error reaches a minimum. This way the model reduces any overfitting over time. In order to find this minimum, the model needs to run a full cycle. I ran the model with 200 epochs with a batch size of 32.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zT4V5gbnnYviDIYRIxK_LQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/540/1*N0euN5lcuQBno8koYyZ8Kg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/694/1*FH0fJzv94boLvTGsr-i5nQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/798/1*omIuFTtndYD-K4iZeYgiag.png" /></figure><p>As you see, the model is overfitting since the Training accuracy is 99% while the test results are around 61%. After 75 epochs the training model still continues to grow and increase accuracy while validation model plateaus. Rerunning the model to stop at 75 epochs resulted in an increase test accuracy of 60%</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*IulnIDerkIMxB6oaKIU7cQ.png" /></figure><h3>L1</h3><p>Using L1 to remove all less important features trains the model to underperform in both the training and test set. This is seen by the training accuracy and test results being having a 4% spread but the results are below 65%. This is showing us that some of the features that were removed from the model have some importance to predicting the wine category based off review. This has been the best model yet, since the training set is not overfitting and the test set is achieving similar results.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*vrX8PdBf3H7N_E7HxpyqYA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/514/1*hBWhvRJkwmwHrmQnAv6-NQ.png" /></figure><h3>L2</h3><p>Running the model by reducing the coefficients of lesser important features closer to zero, continued to have the training model overfit the data. The test results remained close to 60% which is similar to what we have seen in the past models. This did not import our accuracy overall</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3pYfZ3iA-zodhZ_S5_YD1w.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/538/1*OqY9hTlDMRhRvEqlMAZhBg.png" /></figure><h3>Drop Out</h3><p>The Drop out method with a probability of 30% of a neurons being dropped during the cycle, we see our model’s Test Accuracy increase to 64%. The training model is still overfitting since its accuracy is at 100%, but we were able to increase the Test accuracy. One thing to note about Drop out, When running the model on the Test, all neurons will be used since the model wants to have all the neurons that were trained. Increasing the drop out percentage to 50%</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*X3kxCyqDcOLDEN7R4MtgzA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/476/1*wuUZn7QguBLFjvyXsS_iuQ.png" /><figcaption>Drop out model at .3%</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/926/1*DRgUoiWsiHfpZ1I2wHtrwQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/520/1*wS3YH1Wdq8Pqya0MTP6cog.png" /><figcaption>increased Test accuracy by 1%</figcaption></figure><h3>Conclusion</h3><p>The best from the models results came from L2 regularization, where it dropped all features with no significant impact. My model was able to predict the predict the type of wine based off the review with 61% accuracy, with a training accuracy of 65%. Although the drop out model achieved an accuracy of 65%, our training set accuracy was at 99% accuracy, showing that our model was severely over fitting to the training set. Overall these are very good results given that there is a 5% probability one could randomly guess the type of wine based off of a review.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1a79b5e3081f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Date Time Functionality in Python and Pandas]]></title>
            <link>https://medium.com/@jennifer.arty/date-time-functionality-in-python-and-pandas-f2e5393b78d0?source=rss-2dba6326e7b5------2</link>
            <guid isPermaLink="false">https://medium.com/p/f2e5393b78d0</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[time]]></category>
            <category><![CDATA[python]]></category>
            <category><![CDATA[date]]></category>
            <category><![CDATA[pandas]]></category>
            <dc:creator><![CDATA[Jennifer Arty]]></dc:creator>
            <pubDate>Tue, 28 Aug 2018 13:04:57 GMT</pubDate>
            <atom:updated>2018-08-28T13:05:31.299Z</atom:updated>
            <content:encoded><![CDATA[<p>Before learning about datetime functions in Python and Pandas, working with dates was always a tedious process. On projects orassignments I either had to create a new columns to separate the month date and year to sort, or try to make it into a numerical value. Using datetime has made working with these data types a much easier process!</p><p>One way of creating a DateTime datatype is by using the datetime feature in python. Importing datetime you can can strptime() method, which will convert the string to a date time based on the format of the string. Below I have shown 4 examples of how different types of date formats will convert to DateTime. By default, once a date has been converted it will come out as:</p><blockquote><strong>Year-Month-Day Hour:Minute:Seconds</strong></blockquote><p>In order for the string to come out in this format, you need to tell the function what information you are passing through using strtime directive<em>. </em>The format of the directives need to match the format of the datetime string that is being passed through as well as the special chracters.</p><p>For example : Date1 is passing through day of week, Month day and year written out. To account for this we use</p><p>%A =weekday,</p><p>%B =Month,</p><p>%d = day</p><p>%Y= 4-digit year</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/946/1*zX0milYT55P1QYILRZ77MQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*L0Mo3hyGJLcs9l2oT2VP2Q.png" /><figcaption><a href="https://docs.python.org/3/library/datetime.html#strftime-and-strptime-behavior">https://docs.python.org/3/library/datetime.html#strftime-and-strptime-behavior</a></figcaption></figure><p>When calling type on the dt1 variable, we see that it was converted from a string to a datetime data type!</p><p>Pandas also has a DateTime feature built in where you can convert objects or integers into a DateTime data type, which makes analyzing data with continues times a breeze. To convert the column in the data frame, you have to call .to_datetime() method in pandas. Below is an example of the conversion:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/674/1*hqDRY1S_S55b9txtsex_xA.png" /></figure><p>Using an old project on 311 calls in NYC, I was able to convert the date column into a datetime data type and run some analysis on it. With the date in an easy to use form, Pandas makes analyzing data by month, day, hour etc. a breeze. This is done by using Time/Date components for the different timestamps your are interested in.Taking a look at the complaint types of the 311 calls below for noise complaints, I was able to group data by month, hour and day of week. We see that most noise complaints come in the summer months when everyone is outside enjoying the warm weather. When taking a look at the hour, its expected to see the complaints coming in from 10pm to the most coming in at mid-night. Lastly, weekends are the days of week where most of the complaints are coming in.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*U4WXaFGel49KcPpt6t6cqQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1016/1*3rDOctJ0r-t3zIvnpCtUzQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wK1J-SX6gClKdSOxHbv7iQ.png" /><figcaption>Analysizng complaint calls by month, hour, and day of week</figcaption></figure><p>See below for list of different Time/Date Components that can be used</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VXSFlk0sQTNyM-Bx8Kzx5w.png" /><figcaption><a href="https://pandas.pydata.org/pandas-docs/stable/timeseries.html">https://pandas.pydata.org/pandas-docs/stable/timeseries.html</a></figcaption></figure><p>Learning and using DateTime in python and pandas has made a daunting task of tackling dates and time objects a breeze! I am excited to continue to learn and explore all the possible functionality it has to offer from analyzing, visualizing and manipulating time series data.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f2e5393b78d0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Python cheat sheet for NumPy and Pandas.]]></title>
            <link>https://medium.com/@jennifer.arty/python-cheat-sheet-for-numpy-and-pandas-d1ed95815b2e?source=rss-2dba6326e7b5------2</link>
            <guid isPermaLink="false">https://medium.com/p/d1ed95815b2e</guid>
            <category><![CDATA[python]]></category>
            <dc:creator><![CDATA[Jennifer Arty]]></dc:creator>
            <pubDate>Tue, 31 Jul 2018 12:53:51 GMT</pubDate>
            <atom:updated>2018-07-31T16:47:39.069Z</atom:updated>
            <content:encoded><![CDATA[<p>Python has an infinite amount of amazing features that learned to cherish and to love. With many of the functions relating to one another. Last week, I was introduced to the programing library NumPy and Pandas. While learning their individual functions, I noticed the many similarities and differences that they have with one another. Having only been using these libraries and their functions a couple of times, I saw the importance of having a quick cheat of the functions they can both do, in relation to Python List.</p><p>First, its important to know what exactly are NumPy and Pandas libraries:</p><p><strong>NumPy: </strong>a library capable of performing scientific computing in python. NumPy holds multi dimension arrays and data structures. To get access to NumPy and all of its amazing features, you will have to import it using :</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/448/1*ldbd1Zqj8nmXbnYpQhMU9Q.png" /><figcaption>import NumPy into Python</figcaption></figure><p><strong>Pandas: </strong>is an open source library for data and analysis. One of my favorite features is how it can take practically any data file( CSV, Excel, JSON, SQL etc) and make it readable and easy to use in Python. Pandas uses either Series ( one dimensional arrays) or DataFrames(two dimensional data structure). To import Pandas you will use the following code:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/370/1*p-Cxm9FhOqd7OpUHFhgH6A.png" /><figcaption>import Pandas into Python</figcaption></figure><p>You can import any of the above file types using the following code:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/700/1*hBp4ysElkhhv9JEbGIluPg.png" /><figcaption>Using Pandas to import various types of files into Python</figcaption></figure><p>Below in the cheat sheet you will see how to call all major functions for Python List, NumPy arrays and Pandas Series and Pandas DataFrame. I have split them up into the following categories: Slice/index, Length, Sort, Append, Calculations. In the last section of calculations, you will see how use full NumPy and Pandas comes in to calculate anything just with a single function.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4TlRz6qT-TFKm9etPbAu3w.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*PnTdc-aO5fhs6cJ-kLiGSw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lQB8Pn62Bn6ldJss1UnNMA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5bEVqXbklu6dqHlgXuvpOw.png" /></figure><p>Overall you see that there are many shared functions from finding the index of a certain item or calling the length of your data. The key difference is Pandas and NumPy sometimes have different ways of calling it. One of the key importance is to realize what type of data you are dealing with, wether it is a single or two dimension NumPy array or a Pandas Series or Data Frame.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d1ed95815b2e" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>