We are half way through October now and here is a first compilation of papers I found interesting online!
Melinda Has a New Mission: Women in Tech: Melinda Gates is known as the wife of Bill Gates. But she is also an amazing woman who got a degree in Computer science and Economics, as well as an MBA. She also worked actively at Microsoft on multiple big projects before marrying Bill and starting her philanthropist career. It is stated in this post that when she graduated, one student on three graduating from the computer science program was a woman. Today, it is one woman on five. This is a huge step back when it comes to women progress in STEM fields. The post hence encourages others to come up with ideas to tackle this issue and here, we get a candid interview with Melinda herself on the subject.
Coding Like a Girl In this light post on Medium, a few women in the technology industry share their stories on how, even with great qualifications, end up being judged for their physical appearance before anything else.
How to Do Research
Science Isn’t Broken Science and the field of research can only be reliable when the work is well-down and there are concrete cases/data to support the statement being made. And sadly, in our current world of trending titles and sensational news, the idea of being a reliable source is often lost. Here, this post talks about the issue of p-hacking, the idea to of changing interactive data through changing the p-value, just so the data fits with the original hypothesis been made.
The Problem with P-values In the same direction as the previous post, this one also discuss about the accuracy of research being done based on the p-values. Especially, here, we are looking at the danger of false positives results based on a bad choice of p-value.
Open Source for Who It is definitely a valid question to ask and to seek an answer for. The reality is that Open Source is a concept that very few people actually heard of, out of everyone who could actually have access to it. The accessibility of a project or information is not only because it is online or accessible through an API. The author here also mentions the importance of having good documentation, tutorial and ways to link back to this open source project.
Do you still believe Donald Trump can win (August 2016)? Why? Here is another post from Quora where people are discussing the likelihood of Donald Trump winning the upcoming elections. Even if the media wants to make the battle sound equal, the reality is that based on past polling results and the predictive pattern of US elections, there is only about 15% chance that Trump does get elected.
Millennials: Someone You Love May Be Alive Because of Hillary Clinton The media is making a joke out of Hillary Clinton’s efforts to appeal to the Millennium generation. Here, one of her supporter shares her take on the effort and makes us realize just how much she had done for the future generation of the United States.
Data Mining Hacker News: Front vs Back Here, the authors are looking at the historical data of previous posts being posted and reposted on the website Hacker News and how much traffic each of these posts can generate based on variables like time, title, sentiment, and the user. All of this is also supported by Python code.
#TrumpWon? trend vs. reality After the first presidential debate, the Twitter hashtag #TrumpWon was gaining the most attention on the social media and even encouraged the presidential candidate to thank his fervent followers. However, a bit later, it was revealed that this hashtag was mostly supported by users .. in Russia. Now, it is time for our author to look deeper into what really happened. As it turns out, the hashtag was generated in Beltimore and Detroit, USA. It then got more attention in the rest of United States and won some more trendy points in UK and Australia. The author keeps looking into the data more and more and gives a few conclusion. First, the hashtag was automated by actual Twitter users, not bot. However, the efficiency of how fast it started trending is because people of the same hub/group and with influence were using it. There is also no connection with Russia. All in all, this shows that both parties tried hard to twist with the reality such that their party looks better. And that is the reality of social media.
Deep-Fried Data Comparing Machine Learning to deep-frying data is a new way to see how this science has evolved to. In this talk, the speaker discuss the evolution of machine learning and how it can solve many problems out there. On the topic of social media, I especially like this quote:
People face social pressure to abandon their privacy. Being on LinkedIn has become an expected part of getting a job or an apartment. The border patrol wants to look at your social media.
Estimating Delivery Times Most companies offer a service or a product. If that is the case, it implies there is at the end point a customer waiting for that service or product to arrive. Now, machine learning can come in to solve the issue of finding how long waiting time takes. This post takes the example of delivering pizza and makes the reader understand the importance of understanding all the factors before starting the analysis.
Rubik’s Cube Solver An interactive way to solve the Rubik’s Cube with a tutorial, how innovative is that?! Check out the youtube video to see the magic for yourself:
Transit Maps: Apple vs. Google vs. Us Mapping Data is always fascinating to me, because data in a map looks great! In this visually stunning post, the author discuss the science behind transit mapping in the biggest cities in the world. It also approach on how to create a map that can work for all cities.
Three Challenges for Artificial Intelligence in Medicine We all know the power of data science and what it can do in the field of Medicine, a field with so much data waiting to be used! The author suggest that the most important challenges to pay attention to are:
Healthcare as a Label Desert and One-Shot Learning
Deployment and the Outside-In Principle
Regulation and Fear
Generating Faces with Deconvolution Networks Another wonderful article, this one about the science of using programming, data sciences and deconvolution netwrok to generate faces. The idea is to basically input various parameters to the program and wait for it to generate.. interesting results!
Who Funds Quora and Why? Quora was first started off by two past Facebook employees, Adam D’Angelo and Charlie Cheever. Founded in 2009, the site was meant to be an accessible Q&A site.
9 Tricks to Appear Smart in Brainstorming Meetings A fun read with beautiful and funny drawings to back it up! Don’t take it all too seriously, but honestly, these advice can work. 😉