TOC

Author: Thomas Frederikse

jupyter-notebooks/use-cases/pv-yield-forecasting/pv-yield-forecasting.ipynb

@@ -22,7 +22,7 @@
     "[References](#References)\n",
     "\n",
     "\n",
-    "## Hay yield in North Dakota\n",
+    "# Hay yield in North Dakota\n",
     "Agriculture is an important economic sector in the US state of North Dakota, and about 90 percent of its land area is used for farming. \n",
     "From our Planet Satellites, North Dakota can clearly be classified as an agricultural state:\n",
     "\n",
@@ -96,7 +96,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Planetary Variables\n",
+    "# Planetary Variables\n",
     "Over South Dakota, we have a satellite observation of all three Planetary Variables at least every two days. Here, we use the average value of these PV's over the growing season from May to July. \n",
     "\n",
     "### Soil Water Content\n",
@@ -261,7 +261,7 @@
    "metadata": {},
    "source": [
     "\n",
-    "## Statistics\n",
+    "# Statistics\n",
     "\n",
     "We now have three variables that show a very similar pattern as the yield: SWC, LST and VOD. Can we say something about which indicator is the best predictor of yield? We can quantify the performance of each predictor by computing correlation coefficients: bluntly said: the higher the correlation between two time series, the higher the similarity between both. Two widely-used correlation coefficients are Spearman $R$ and Pearson $R$. \n",
     "\n",
@@ -404,7 +404,7 @@
     "\n",
     "The correlation between VOD and SWC/LST is much smaller. This means that changes in VOD often differ from changes in LST/SWC and therefore, VOD might add additional predictive information on yield on top of SWC and LST.\n",
     "\n",
-    "## Forecasting yield using SWC, LST, and VOD\n",
+    "# Forecasting yield using SWC, LST, and VOD\n",
     "We have now learned that SWC, LST, and VOD correlate with the yield. Now we want to know whether we can make a more accurate yield prediction using a combination of these three Planetary Variables. Here, we are going to use a simple regression model to find out. We want to solve the following model:\n",
     "\n",
     "$$\n",
@@ -924,15 +924,15 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Conclusions\n",
+    "# Conclusions\n",
     "We have made a forecast of the hay yield in North Dakota for the year 2024 using a simple regression with three Planetary Variables (SWC, LST, and VOD). While this analysis is rather simple and lacks the rigorous testing needed to assess the reliability of the regression, it does show that Planetary Variables are a good indicator of crop growth: all three PV's show a clear correlation with the annual yield. "
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## References\n",
+    "# References\n",
     "\n",
     "### Data sources\n",
     "- [USDA yield data](https://www.nass.usda.gov/Data_and_Statistics/)\n",