Move files

src/data/projects/forex.ts

@@ -0,0 +1,36 @@
+import forexPosterWebP from '../media/forex.jpg?format=webp';
+import forexPosterJpeg from '../media/forex.jpg?format=jpg';
+import forexMp4 from '../media/mp4/forex.mp4';
+import forexWebM from '../media/webm/forex.webm';
+
+import { Video } from '../../page/basics/video/video';
+import { TimelineElementParameters } from '../../page/timeline/timeline-element/timeline-element';
+
+export const forexTimelineElement: TimelineElementParameters = {
+  title: `Predicting foreign exchange rates`,
+  date: `2019 autumn`,
+  figure: new Video({
+    posterWebP: forexPosterWebP,
+    posterJpeg: forexPosterJpeg,
+    mp4: forexMp4,
+    webm: forexWebM,
+    invertButton: true,
+  }),
+  description: `
+    From the animation, we can see that my implementation does a somewhat acceptable job at
+    predicting (blue graph) the EUR/USD rates (green graph).
+  `,
+  more: [
+    `
+      In a nutshell, the algorithm (written in Python using NumPy, SciPy, and Flask)
+      predicts in the frequency domain. The steps are the following: smoothing the input values, 
+      differentiating, applying a short-time Fourier-transformation with overlapped (and Hanning-windowed) windows,
+      extrapolating and then applying the inverse of these transformations to the resulting values.
+    `,
+    `
+      Of course, there is still plenty of room for improvement, but even with this simple algorithm
+      a mostly profitable trading strategy is viable. In my free time I may put more work into it.
+    `,
+  ],
+  links: [],
+};