So, it turns out my idea of posting every week or two has already gone right out of the window. I know full well that all of you, yes, all zero of you reading this, have been missing me dearly. You are probably wondering where I’ve been and what I’ve been doing… right?! Well, in my super exciting life, I’ve been busy paper writing…. okay, so not that exciting. However, said paper recently got accepted into the journal ‘Frontiers in Plant Science’ titled ‘A deep learning method for fully automatic stomatal morphometry and maximal conductance estimation’. Simply put; I developed a convolutional neural network to identify stomata – pores on a leaf that allow the exchange of gases between the atmosphere and the plant – in images obtained via a microscope, performed morphometric analysis using the detections from the network and some image processing methods, and estimated the anatomical maximal conductance rate. It is an improvement over the majority of stomata literature that simply focus on identifying stomata, via bounding boxes, to provide counts which in turn can lead to a prediction of density. The method I proposed uses pixel wise segmentation which supports the quantification of stomatal traits.
This leads me to my rant. Important disclaimer: The next paragraph contains my endless boring moaning, skip to avoid such content.
<Rant> The paper was initially submitted to New Phytologist and was rejected by one reviewer based on the novelty, which would be fine… if this was the case. They went on to claim that they could have done this but hadn’t.… so, why didn’t they? And why does this make it not novel? There are a lot of things wrong with academia and the everyman-woman for themselves thing certainly does not help science progress. It is disappointing to see the lack of progress made when improved collaboration could accomplish so much more. Even more so given the current climate conditions and lack of changes by governments. </Rant>
Moving on to the rest of the month, and on a brighter note, the kind people at Polygon.IO, particularly Jack Bell who was incredibly helpful (and also owns an adorable German Shepard) granted me access to an academic license of their API. If you haven’t heard of Polygon, it allows you to access to real time stock market data. My use for Polygon is purely academic but, what it offers, will undoubtedly help you day traders out there. So far, I have created a simple web interface for the data which allows real time viewing of stock market data, news scraping and data visualisation. Within the next week or two I will be including a tutorial here for those of you considering using Polygon or are generally interested in stock market data. Note that they do offer a free license, but it does have limitations. and no, unfortunately I’m not getting paid for advertising.
With respect to research, common literature in relation to stock markets look at predicting the price on a daily basis over a given period of time. However, my research focuses on predicting whether the stock will move up or down in price on the next day based on intraday trading data and catalysts. Technically speaking if you were a day trader predicting where the stock will move the next day provides a significant edge. Predicting longer term trends can suffer immediately if the first prediction is wrong. I will keep you updated on this.
That’s all for this time. I’ll update date again shortly with some tutorials that you will be able to find here.
^^ back to top ^^