Taking this a step further, the perk of having a spike in new users is that their trial periods more-or-less overlap. Usually when a new user signs up, they will poke around the product and maybe send in a bug report, a feature request or another piece of feedback. If I reply and address their feedback by putting in a fix or shipping their requested feature within days, more often than not, they end up converting into a happy customer.

Something that I’m glad I didn’t spend the time to automate at all was the referral program. If an existing user refers a new user and the new user subscribes to a plan, then both users will receive credit for 1 month. In the end, the referral program only brought on 11 extra users. It was not a pain at all to manually process rewards for those who did. If I had spent time conjuring up all the edge cases and automating this process, it would not have been worth the time investment.

But I’ve noticed that my working cadence is, I’ll be deep into a major feature or maybe a group of improvements that I’d like to make on a particular feature and I’ll work on it for three days. Once I deploy and I notify my users, it’s like this wave comes crashing down and I lose the motivation to work a little bit. I think that’s my body telling me, hey, I need a break. So my schedule is basically work really hard for three to four days and then take a day, day and a half off. I find that that works well for me.

On the subject of talking to your friends and family and trying to get accurate feedback instead of having them telling you what you want to hear so you feel better, there’s a whole book called The Mom Test that I highly recommend.

But I think you also bring up a good point that a lot of people feel that the personal finance and budgeting app space is oversaturated. I don’t think that at all, because I feel that personal finance is a very personal thing. Everyone has their own way that they like to do things. I don't think that this is something where you’re going to find over 50% of people share the same budgeting philosophy. So I feel that if you have a preferred way of doing something, you could probably find a cohort out there that feels the same way.

The following diagram shows the various components of a template.

The template author defines a template, which describes the data Mixer dispatches to adapters, and the interface that the adapter must implement to process that data.

Packages

Any

Also note that if a scalar message field is set to its default, the value will not be serialized on the wire.

Note that for scalar message fields, once a message is parsed there's no way of telling whether a field was explicitly set to the default value (for example whether a boolean was set to false) or just not set at all

Scalar Value Types

When a new array is created by indexing with an array of integers, the new array has the same shape than the array of integers:

>>> a = np.arange(10, 0, -1)
>>> idx = np.array([[3, 4], [9, 7]])
>>> a[idx]
array([[7, 6],
       [1, 3]])

A slicing operation creates a view on the original array, which is just a way of accessing array data.

a[a % 3 == 0] = -1

NumPy arrays can be indexed with slices, but also with boolean or integer arrays (masks). This method is called fancy indexing. It creates copies not views.

Worked example: Prime number sieve

is_prime = np.ones(101, dtype=bool)
is_prime[:2] = False
n_max = int(np.sqrt(len(is_prime) - 1))
for j in range(2, n_max + 1):
    if is_prime[j]:
        is_prime[j*j::j] = False
np.nonzero(is_prime)

np.arange(6) + np.arange(0, 51, 10)[:, np.newaxis]

Exercise: Tiling for array creation

a = np.array([[4, 3], [2, 1]])
np.tile(a, (2, 3))

Exercise: Array creation

a = np.ones((4, 4), dtype=np.int32)
a[2, 3] = 2
a[3, 1] = 6

# I've look up the solution in 
# https://github.com/pv/numpy-tutorial/blob/master/solutions/1_1_array_creation.py
b = np.zeros((6, 5))
b[1:] = np.diag(np.arange(2, 7))

Try the different flavours of slicing, using start, end and step: starting from a linspace, try to obtain odd numbers counting backwards, and even numbers counting forwards.

> a = np.arange(10)
> a[::-1]
array([9, 8, 7, 6, 5, 4, 3, 2, 1, 0])
> a[9:-1:-1]
array([9, 8, 7, 6, 5, 4, 3, 2, 1, 0])
> a[8:0:-1]
array([8, 7, 6, 5, 4, 3, 2, 1])

A small illustrated summary of NumPy indexing and slicing...

Plot some simple arrays: a cosine as a function of time and a 2D matrix.

import numpy as np
import matplotlib.pyplot as plt

# Compute the x and y coordinates for points on sine and cosine curves
x = np.arange(0, 3 * np.pi, 0.1)
y_sin = np.sin(x)
y_cos = np.cos(x)

# Plot the points using matplotlib
plt.plot(x, y_sin)
plt.plot(x, y_cos)
plt.xlabel('x axis label')
plt.ylabel('y axis label')
plt.title('Sine and Cosine')
plt.legend(['Sine', 'Cosine'])
plt.show()

Look at the function np.empty. What does it do? When might this be useful?