diff --git a/README.md b/README.md
index 79b069d..15afd57 100644
--- a/README.md
+++ b/README.md
@@ -27,7 +27,7 @@ as implemented by `scipy`
       the fit parameters are: FitParams(a=2.2195448757323435, loc=0.0)
       the negative log likelihood is: 7.86098064513523 
     
-    ![image](fits.png)
+![image](fits.png)
 
 ## Notes
 The goodness of fit is determined here using the maximum likehood approach. The fit itself is done by [`scipy.stats.fit`](https://docs.scipy.org/doc/scipy/reference/generated/scipy.stats.fit.html#scipy.stats.fit) which optimises for the parameters that maximise the likelyhood estimate. Note, `scipy.stats.fit` is a new feature introduced in scipy 1.9.0 that allows seamless fitting both discrete and continuous distributions.
@@ -35,9 +35,9 @@ The goodness of fit is determined here using the maximum likehood approach. The
 The goodness of fit metric shown here is the negative log of the [Probability Mass Function (PMF)](https://docs.scipy.org/doc/scipy/tutorial/stats/discrete.html?highlight=fit#probability-mass-function-pmf).
 
 ## Bounds
-Fitting requires some rough information about parameters bounds or a guess value to start from. Here, I hard-coded very rough bounds that could very well fail for a variety of data edge cases. These bounds were tested on datasets sampled from uniform distributions. 
+Fitting requires some rough information about parameters bounds or a guess value to start from. Here, I hardcoded very rough bounds that could very well fail for a variety of data edge cases. These bounds were tested on datasets sampled from uniform distributions. 
 
 ## Future work
 The script could easily be extended to fit any other [scipy discrete distribution](https://docs.scipy.org/doc/scipy/tutorial/stats/discrete.html?highlight=maximum%20likelihood#discrete-distributions-in-scipy-stats). If additional distributions are required one would have to manually write the distribution function, and then manually find the parameters that maximise the PMF. 
 
-Ideally the choise of distributions would not be hardcoded and the user could choose based on looking at their data what distributions to try. 
+Ideally, the choice of distributions would not be hardcoded and the user could choose based on looking at their data what distributions to try.