DEV Community: CristinaTRico

Commitment Phobia

CristinaTRico — Sat, 01 Feb 2020 16:50:46 +0000

While completing group projects at The Flatiron school I had some serious commitment phobia when it came to GitHub. I was so afraid my changes were not being saved, or that something would go wrong and my work would be ruined.

Over the last week though I've worked through that fear by making quite a few commits.

Last week I blogged about how to create a repository from the command line and today I will take you through the steps to add files to your repository and update already existing files in your repository.

Step 1: Adding/Staging a new file to your repository

Using the command git add you can stage local files to an existing repository. staging a file means preparing the file for a commit.

git add. will stage all local directory files. You can stage individual files from your directory by using git add filename.ext instead of git add.

Step 2: Commit and Comment on Changes

Using the command git commit -m "comment" allows you to commit the file or files you staged in step 1. the -m allows you to leave a comment on the changes made. It is important to comment on your commits so that you leave a trail of changes to your files.

Step 3: Push the changes to the repository

Using the command git push you will push your changes to the repository.

It really is easy as 1-2-3.

Time to Git to it - Creating a repo from the command line

CristinaTRico — Sat, 25 Jan 2020 18:54:33 +0000

Coming from Flatiron there is a checklist of things I'm required to do each week during my job search. Although I've already been doing them I have not been tracking them until recently. One of the things being tracked is one of the basics I struggled with during my time at Flatiron: using Git and Github. Knowing that I will be required to make a minimum of 5 Git commits per week I've decided to review some Git basics.

Generally, I create a new repository on the GitHub site, but did you know you can actually do this from the terminal?
To begin, navigate to the directory you wish for this new repo to live. If you have access to the curl command which comes built-in with Mac, you can create a repository from scratch by making a call to the GitHub API.

Send a post request using the following:
curl -u "YourGitHubUsername" http://api.github.com/user/repos -d '{"name":"Name of Your Repository"}'

When you enter, you will be asked for your GitHub password.

You've just created a new repository from the command line!

Boxplots +

CristinaTRico — Thu, 08 Aug 2019 17:05:55 +0000

Having spent the last 6 years fielding the question, "When are we ever going to use this in real life?" I'm excited to approach boxplots with a different audience in mind.

The Basics:

The boxplot is a standardized method used to display the distribution of data based on a 5 number summary: minimum, quartile 1 median, quartile 2, and maximum.

IQR: interquartile range, calculated as Q3 - Q1
Minimum: Q1 - 1.5 * IQR
Quartile 1: 25th percentile
Median: 50th percentile center value of the data's full range
Quartile 3: 75th percentile
Maximum: Q3 + 1.5 * IQR

Image Source

Beyond the summary lines, boxplots allow you to view outliers, symmetry, tightness of grouping, and any skews your data may have. Boxplots are best used when you want more information than just the central measures of tendency as they make dispersion and variability visible.

Group Comparisons:

Boxplots work their magic when you are comparing many groups of data sets. The visible 5 summary statistics make similarities and differences between groups evident. For example, the following plot was created while exploring trends in movie data. The release dates of movies ranging from 2006 - 2018 were grouped by release week and plotted against %ROI. The aim was to determine if there were weeks that were favorable to others in terms of profitability.

Here the visual effect was enhanced by utilizing a colormap. Based on the percentile of data above or below the break-even point, the boxplots are shaded in a gradient or red or blue. The visual allows for quick insight, directing further exploration.

Extensions:
Despite the depth of information, it could be useful to have a more direct means of comparison between plots. Focusing on a subset of the release week data I wanted to implement notched boxplots and a swarm overlay.

Notched Boxplots:
The notched boxplot displays a confidence interval around the median, generally a 95% confidence interval. If the notches in the boxplots being compared do not overlap then there is evidence that the medians are significantly different. I chose a few weeks where I expected some medians to be significantly different and some not to be.

The notches help to display that the medians for the weeks of May 17th and May 31st are significantly different than those for the weeks of April 19th and April 26th.

Swarm Overlay Boxplots:
One downside to boxplots is being unable to see all observations that are underlying the distribution of the plot. A swarm plot overlay not only allows us to see all observations but ensures data points are a minimum distance away from one another to over overlapping. The swarm overlay gives a better representation of the distribution of values.

To get the full effect of the swarm overlay, be sure to change the transparency level of your box plots or your points will be hidden behind the box itself.

Why I Choose to Study Data Science

CristinaTRico — Thu, 25 Jul 2019 14:38:32 +0000

I can approach this question similar to how one could approach data, or so I am learning. There are numerous ways and they vary in readability, depth, and finesse. So I will focus on readability, delve as deep as is necessary, and remind you that I am a beginner, both in blogs and in data. Overall, I choose to study data science because I want to be learning, I want to be challenged, I want to be proud of myself.

If I could have stayed in college forever I would have. My family would say that I tried. When I think back to the years I was most happy and fulfilled it is easy to pinpoint the years I was in college. The feeling of accomplishment gained from learning and growing daily is something I took for granted and sometimes even cursed while I was there.

Having graduated, my feeling of accomplishment from learning had to come from new job experiences or personal interests. Learning experiences in my previous jobs tended to dimish with time, some sooner than others. Interest in my work would die soon after I realized my time of learning was over. With the ever-increasing growth in data however, there will be no shortage of questions to ask, skills to learn, or methods to master. The ability to continue to question, to continue honing my craft, and possibly contribute to the field of data science is what speaks most to my desire to forever be learning.

I will face challenges as I pursue a career in data science. Of this, I am certain, because it is an everyday occurrence already. But I've never wanted to pursue anything that was not a challenge. I was a first-generation college graduate, I completed a summer research program at the University of Maryland, and I was admitted into a competitive Ph.D. program in Economics, and I taught 13-year-old girls Algebra, none of these things came easily. It was the challenge that drove me, that kept me interested, and made it feel worth doing. Although I am just beginning my journey I have already faced and surpassed challenges. These challenges have just been a missed colon, a typo, forgotten syntax, or not understanding what people around me were talking about. Regardless of how simple the challenges have been so far, they did their job, they made me doubt my ability. Surpassing them, though, strengthened my determination to rise to the challenges.

As I mentioned above, if I could have stayed in college forever I would have. So many find it confusing when I tell them that I dropped out of my Ph.D. program only two weeks in. This was 7 years ago now, and yet it crosses my mind almost daily. I had put so much time and effort into getting admitted to the program but once I got there I lost my nerve.

My classmates all had a higher academic background than me. They had already completed a master's or Ph.D. in another subject. I was the only student that had come in relying solely on a bachelor's degree. I was intimidated. Feeling intimidated coupled with the doubt that I chose the wrong path in pursuing economics instead of statistics consumed my every thought. Between the intimidation and the doubt, I chose to focus on the doubt. I convinced myself that I had chosen the wrong path. I convinced myself I had no passion for economics. I convinced myself I would regret continuing in my program. I convinced myself I wouldn't have felt this way had I chosen the statistics program instead. It was easier to say I had made the wrong choice than admit that I was an imposter. The truth is I dropped out of my program because I felt like an imposter. I've been looking for a way to prove confront those feelings again and instead persist past them. I can't think of a better field than data science to do exactly that. The ability to investigate, answer questions, and then present those findings, will not assuage me feeling like an imposter; but, continuing to do so regardless of those feelings eventually will.

I could have used this blog entry to talk about my passion for data, statistical sleuthing, or the amazing job market that will be open to me, but for me, this is more personal than that.