diff --git a/07.2_Algorithms.ipynb b/07.2_Algorithms.ipynb
new file mode 100644
index 0000000..8e69512
--- /dev/null
+++ b/07.2_Algorithms.ipynb
@@ -0,0 +1,813 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "import matplotlib.pyplot as plt\n",
+    "\n",
+    "%matplotlib inline\n",
+    "#plt.style.use('ggplot')\n",
+    "%precision 4\n",
+    "#np.set_printoptions(suppress=True)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Interpolation: \n",
+    "\n",
+    "`scipy.interpolate` is useful for fitting a function from experimental data and thus evaluating points where no measure exists. The module is based on the FITPACK Fortran subroutines."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# the actual measurements\n",
+    "measured_time = np.linspace(0, 1, 10)\n",
+    "noise = (np.random.random(10)*2 - 1) * 1e-1\n",
+    "measures = np.sin(2 * np.pi * measured_time) + noise\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from scipy.interpolate import interp1d\n",
+    "# linear interpolation\n",
+    "linear_interp = interp1d(measured_time, measures)\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "The kind of interpolation can be specified: linear is the default;  *zero*, *slinear*, *quadratic* and *cubic* refer to a [spline](https://en.wikipedia.org/wiki/Spline) interpolation of zeroth, first, second or third order; *previous* and *next* simply return the previous or next value of the point."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# cubic interpolation\n",
+    "cubic_interp = interp1d(measured_time, measures, kind='cubic')\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# denser interval to plot the interpolating function\n",
+    "interpolation_time = np.linspace(0, 1, 50)\n",
+    "\n",
+    "plt.figure(figsize=(6, 4))\n",
+    "plt.plot(measured_time, measures, 'o', ms=6, label='measures')\n",
+    "\n",
+    "linear_results = linear_interp(interpolation_time)\n",
+    "plt.plot(interpolation_time, linear_results, label='linear interp')\n",
+    "\n",
+    "cubic_results = cubic_interp(interpolation_time)\n",
+    "plt.plot(interpolation_time, cubic_results, label='cubic interp')\n",
+    "plt.legend()\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# checking the punctual results\n",
+    "x=1/4.\n",
+    "print (\"sin(x):\",np.sin(2 * np.pi * x), \n",
+    "       \"linear prediction:\",linear_interp(x),\n",
+    "       \"cubic prediction:\", cubic_interp(x))\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Optimization and fit: `scipy.optimize`\n",
+    "\n",
+    "Optimization is the problem of finding a numerical solution to a minimization or equality.\n",
+    "\n",
+    "The `scipy.optimize` module provides algorithms for function minimization (scalar or multi-dimensional), curve fitting and root finding."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from scipy import optimize"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Curve Fitting\n",
+    "\n",
+    "Let's take some data, sampled from a sine distribution and affected by random noise"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "x_data = np.linspace(-5, 5, num=50)\n",
+    "y_data = 2.9 * np.sin(1.5 * x_data) + np.random.normal(size=50)\n",
+    "\n",
+    "import matplotlib.pyplot as plt\n",
+    "plt.figure(figsize=(6, 4))\n",
+    "plt.scatter(x_data, y_data)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "If we *known* that the data is sampled from a sine function and we just need to find its parameters, in this case amplitude and frequency, we can do that by least squares curve fitting. "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def f(x, A, nu):\n",
+    "    return A * np.sin(nu * x)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# fit by means of curve_fit\n",
+    "# p0 are the guessed values for the parameters\n",
+    "params, params_covariance = optimize.curve_fit(f, x_data, y_data, p0=[2, 2])\n",
+    "print(params)\n",
+    "print (params_covariance)\n",
+    "\n",
+    "# get the full output\n",
+    "#optimize.curve_fit(f, x_data, y_data, p0=[2, 2], full_output=True)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "As every fitting problem, this transaltes into a minimization problem: the function `f(x_data; params)-y_data` is minimized"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plt.scatter(x_data, y_data, label='Data')\n",
+    "plt.plot(x_data, test_func(x_data, params[0], params[1]),\n",
+    "         label='Fitted function')\n",
+    "\n",
+    "plt.legend(loc='best')"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "##  Roots of scalar and functions\n",
+    "\n",
+    "Several methods are there for $f(x): {\\rm I\\!R} \\to {\\rm I\\!R}$, e.g. [Newton-Raphson](https://en.wikipedia.org/wiki/Newton%27s_method) or Brent (a safer version of the secant method, which combines a bracketing strategy with a parabolic approximation)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def f(x):\n",
+    "    return x**3-3*x+1\n",
+    "\n",
+    "x = np.linspace(-3,3,100)\n",
+    "plt.axhline(0)\n",
+    "plt.plot(x, f(x),'r-');"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from scipy.optimize import brentq\n",
+    "brentq(f, -3, 0), brentq(f, 0, 1), brentq(f, 1,3)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from scipy.optimize import newton\n",
+    "newton(f, -3), newton(f, 0), newton(f, 3)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "In case of $f(x): {\\rm I\\!R}^2 \\to {\\rm I\\!R}$, the `optimize.root` is available, also implementing several methods (MINPACK's hybrd being the default)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from scipy.optimize import root, fsolve\n",
+    "\n",
+    "def f(x):\n",
+    "    return [x[1] - 3*x[0]*(x[0]+1)*(x[0]-1),\n",
+    "            .25*x[0]**2 + x[1]**2 - 1]\n",
+    "\n",
+    "solutions = root(f, (0.5, 0.5))\n",
+    "print (solutions)\n",
+    "\n",
+    "print (f(solutions.x))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Minimization of a function\n",
+    "\n",
+    "The `minimize_scalar` function will find the minimum, and can also be told to search within given bounds. By default, it uses the Brent algorithm."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def f(x):\n",
+    "    return x**4 + 3*(x-2)**3 - 15*(x)**2 + 1\n",
+    "\n",
+    "x = np.linspace(-8, 5, 100)\n",
+    "plt.plot(x, f(x));"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "optimize.minimize_scalar(f, method='Brent')\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "optimize.minimize_scalar(f, method='bounded', bounds=[0, 6])\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Local and global minima"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def f(x, offset):\n",
+    "    return -np.sinc(x-offset)\n",
+    "\n",
+    "x = np.linspace(-20, 20, 100)\n",
+    "plt.plot(x, f(x, 5));"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# note how additional function arguments are passed in\n",
+    "solution = optimize.minimize_scalar(f, args=(5,))\n",
+    "solution"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Optimization results are packed accordingly to  this [format](https://docs.scipy.org/doc/scipy/reference/generated/scipy.optimize.OptimizeResult.html#scipy.optimize.OptimizeResult)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plt.plot(x, f(x, 5))\n",
+    "plt.axvline(sol.x,color='b')"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "When the function oscillates a lot, often the global minimum is not found. To cope with that, several ranges needs to be explored"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "lower = np.random.uniform(-20, 20, 100)\n",
+    "upper = lower + 1\n",
+    "solutions = [optimize.minimize_scalar(f, args=(5,), bracket=(l, u)) for (l, u) in zip(lower, upper)]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "idx = np.argmin([solution.fun for solution in solutions])\n",
+    "solution = solutions[idx]\n",
+    "\n",
+    "plt.plot(x, f(x, 5))\n",
+    "plt.axvline(solution.x, color='b');"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Alternatively a dedicated method is available"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from scipy.optimize import basinhopping\n",
+    "\n",
+    "x0 = 0\n",
+    "solution = basinhopping(f, x0, stepsize=1, minimizer_kwargs={'args': (5,)})\n",
+    "solution"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plt.plot(x, f(x, 5))\n",
+    "plt.axvline(solution.x, color='b');"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Minimization of f(x): {\\rm I\\!R}^n \\to {\\rm I\\!R}$ is a whole field of study per se. You should have seen it treated in details in other courses.\n",
+    "\n",
+    "# Statistics and random numbers: `scipy.stats`\n",
+    "\n",
+    "The module scipy.stats contains statistical tools and probabilistic descriptions of random processes. Random number generators for various random process can be found in numpy.random.\n",
+    "\n",
+    "Given observations of a random process, their histogram is an estimator of the random process’s PDF (probability density function):"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Sample from a normal distribution using numpy's random number generator\n",
+    "samples = np.random.normal(size=10000)\n",
+    "\n",
+    "# Compute a histogram of the sample\n",
+    "bins = np.linspace(-5, 5, 60)\n",
+    "\n",
+    "# Compute the PDF on the bin centers from scipy distribution object\n",
+    "from scipy import stats\n",
+    "bin_centers = 0.5*(bins[1:] + bins[:-1])\n",
+    "pdf = stats.norm.pdf(bin_centers)\n",
+    "\n",
+    "from matplotlib import pyplot as plt\n",
+    "plt.figure(figsize=(6, 4))\n",
+    "_,_,_ = plt.hist(samples, bins, label=\"Histogram of samples\", density=True)\n",
+    "plt.plot(bin_centers, pdf, label=\"PDF\")\n",
+    "plt.legend()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "If we know that the random process belongs to a given family of random processes, such as normal processes, we can do a maximum-likelihood fit of the observations to estimate the parameters of the underlying distribution. Here we fit a normal process to the observed data:\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "mean, std = stats.norm.fit(samples)\n",
+    "print (mean, std)  "
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Statistical tests\n",
+    "\n",
+    "A statistical test is a decision indicator. For instance, if we have two sets of observations, that we assume are generated from Gaussian processes, we can use a [T-test](https://en.wikipedia.org/wiki/Student%27s_t-test) to decide whether the means of two sets of observations are significantly different:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Generates 2 sets of observations\n",
+    "samples1 = np.random.normal(0, size=100)\n",
+    "samples2 = np.random.normal(0.1, size=100)\n",
+    "\n",
+    "# Compute a histogram of the sample\n",
+    "bins = np.linspace(-4, 4, 30)\n",
+    "histogram1, bins = np.histogram(samples1, bins=bins, density=True)\n",
+    "histogram2, bins = np.histogram(samples2, bins=bins, density=True)\n",
+    "\n",
+    "plt.figure(figsize=(6, 4))\n",
+    "plt.hist(samples1, bins=bins, density=True, label=\"Samples 1\")\n",
+    "plt.hist(samples2, bins=bins, density=True, label=\"Samples 2\")\n",
+    "plt.legend(loc='best')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "stats.ttest_ind(samples1, samples2)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# most of the test statistics are available \n",
+    "stats.chisquare(samples1,samples2)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Numerical integration: `scipy.integrate`\n",
+    "\n",
+    "### Function integrals\n",
+    "\n",
+    "The most generic integration routine is `scipy.integrate.quad()`, which integrate from a to b (possibly infinite interval) using a technique from the Fortran library QUADPACK.\n",
+    "\n",
+    " Let's compute $\\int_0^{\\pi/2} \\sin{\\theta}d\\theta$:\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from scipy.integrate import quad, quadrature\n",
+    "res, err = quad(np.sin, 0, np.pi/2)\n",
+    "\n",
+    "# An alternative is the gaussian quadrature\n",
+    "#res, err = quadrature(np.sin, 0, np.pi/2)\n",
+    "\n",
+    "print (res, err)\n",
+    "print (np.allclose(res, 1))   # res is the result, is should be close to 1\n",
+    "\n",
+    "print (np.allclose(err, 1 - res))  # err is an estimate of the err"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "`scipy.integrate.nquad()` generalize to multiple integration variables"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from scipy.integrate import nquad\n",
+    "func = lambda x0,x1,x2,x3 : x0**2 + x1*x2 - x3**3 + np.sin(x0) + (\n",
+    "                                1 if (x0-.2*x3-.5-.25*x1>0) else 0)\n",
+    "points = [[lambda x1,x2,x3 : 0.2*x3 + 0.5 + 0.25*x1], [], [], []]\n",
+    "def opts0(*args, **kwargs):\n",
+    "    return {'points':[0.2*args[2] + 0.5 + 0.25*args[0]]} \n",
+    "\n",
+    "result, abserr, out = nquad(func, [[0,1], [-1,1], [.13,.8], [-.15,1]], opts=[opts0,{},{},{}], full_output=True)\n",
+    "\n",
+    "print (result, abserr)\n",
+    "print (out)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Integrating differential equations\n",
+    "\n",
+    "`scipy.integrate` also features routines for integrating Ordinary Differential Equations (ODE). In particular, `scipy.integrate.odeint()` solves ODE of the form:\n",
+    "\n",
+    "```python\n",
+    "dy/dt = rhs(y1, y2, .., t0,...)\n",
+    "```\n",
+    "\n",
+    "As an introduction, let us solve the ODE $\\frac{dy}{dt} = -2 y$ between $t = 0 \\dots 4$, with the initial condition $y(t=0) = 1$. First the function computing the derivative of the position needs to be defined:\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def calc_derivative(ypos, time):\n",
+    "    return -2 * ypos"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Then, to compute y as a function of time:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from scipy.integrate import odeint\n",
+    "time_vec = np.linspace(0, 4, 40)\n",
+    "y = odeint(calc_derivative, y0=1, t=time_vec)\n",
+    "\n",
+    "\n",
+    "plt.figure(figsize=(4, 3))\n",
+    "plt.plot(time_vec, y)\n",
+    "plt.xlabel('t: Time')\n",
+    "plt.ylabel('y: Position')\n",
+    "plt.tight_layout()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Let us integrate a more complex ODE: a damped spring-mass oscillator. The position of a mass attached to a spring obeys the 2nd order ODE $y'' + 2 \\varepsilon \\omega_0  y' + \\omega_0^2 y = 0$ with $\\omega_0^2 = k/m$ with $k$ the spring constant, $m$ the mass and $\\varepsilon = c/(2 m \\omega_0)$ with $c$ the damping coefficient. We set:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "mass = 0.5  # kg\n",
+    "kspring = 4  # N/m\n",
+    "cviscous = 0.4  # N s/m\n",
+    "\n",
+    "# and thus\n",
+    "eps = cviscous / (2 * mass * np.sqrt(kspring/mass))\n",
+    "omega = np.sqrt(kspring / mass)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "For `odeint()`, the 2nd order equation needs to be transformed in a system of two first-order equations for the vector $Y = (y, y')$: the function computes the velocity and acceleration:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def calc_derivative(yvec, time, eps, omega):\n",
+    "    return (yvec[1], -eps * omega * yvec[1] - omega **2 * yvec[0])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "time_vec = np.linspace(0, 10, 100)\n",
+    "yinit = (1, 0)\n",
+    "yarr = odeint(calc_derivative, yinit, time_vec, args=(eps, omega))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plt.figure(figsize=(4, 3))\n",
+    "plt.plot(time_vec, yarr[:, 0], label='y')\n",
+    "plt.plot(time_vec, yarr[:, 1], label=\"y'\")\n",
+    "plt.legend(loc='best')"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Fast Fourier Trasform\n",
+    "\n",
+    "The `scipy.fftpack` module computes fast Fourier transforms (FFTs) and offers utilities to handle them. The main functions are:\n",
+    "\n",
+    "* `scipy.fftpack.fft()` to compute the FFT\n",
+    "* `scipy.fftpack.fftfreq()` to generate the sampling frequencies\n",
+    "* `scipy.fftpack.ifft()` computes the inverse FFT, from frequency space to signal space\n",
+    "\n",
+    "Let's generate a noisy signal:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "from scipy import fftpack\n",
+    "from matplotlib import pyplot as plt\n",
+    "\n",
+    "# Seed the random number generator\n",
+    "np.random.seed(1234)\n",
+    "\n",
+    "time_step = 0.02\n",
+    "period = 5.\n",
+    "\n",
+    "time_vec = np.arange(0, 20, time_step)\n",
+    "sig = (np.sin(2 * np.pi / period * time_vec)\n",
+    "       + 0.5 * np.random.randn(time_vec.size))\n",
+    "\n",
+    "plt.figure(figsize=(6, 5))\n",
+    "plt.plot(time_vec, sig, label='Original signal')\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Let's compute and plot the power spectrum"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# The FFT of the signal\n",
+    "sig_fft = fftpack.fft(sig)\n",
+    "\n",
+    "# And the power (sig_fft is of complex dtype)\n",
+    "power = np.abs(sig_fft)\n",
+    "\n",
+    "# The corresponding frequencies\n",
+    "sample_freq = fftpack.fftfreq(sig.size, d=time_step)\n",
+    "\n",
+    "# Plot the FFT power\n",
+    "plt.figure(figsize=(6, 5))\n",
+    "plt.plot(sample_freq, power)\n",
+    "plt.xlabel('Frequency [Hz]')\n",
+    "plt.ylabel('plower')\n",
+    "\n",
+    "# Find the peak frequency: we can focus on only the positive frequencies\n",
+    "pos_mask = np.where(sample_freq > 0)\n",
+    "freqs = sample_freq[pos_mask]\n",
+    "peak_freq = freqs[power[pos_mask].argmax()]\n",
+    "\n",
+    "# Check that it does indeed correspond to the frequency that we generate\n",
+    "# the signal with\n",
+    "np.allclose(peak_freq, 1./period)\n",
+    "\n",
+    "# An inner plot to show the peak frequency\n",
+    "axes = plt.axes([0.55, 0.3, 0.3, 0.5])\n",
+    "plt.title('Peak frequency')\n",
+    "plt.plot(freqs[:8], power[:8])\n",
+    "plt.setp(axes, yticks=[])\n",
+    "\n",
+    "# scipy.signal.find_peaks_cwt can also be used for more advanced\n",
+    "# peak detection"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "high_freq_fft = sig_fft.copy()\n",
+    "high_freq_fft[np.abs(sample_freq) > peak_freq] = 0\n",
+    "filtered_sig = fftpack.ifft(high_freq_fft)\n",
+    "\n",
+    "plt.figure(figsize=(6, 5))\n",
+    "plt.plot(time_vec, sig, label='Original signal')\n",
+    "plt.plot(time_vec, filtered_sig, linewidth=3, label='Filtered signal')\n",
+    "plt.xlabel('Time [s]')\n",
+    "plt.ylabel('Amplitude')\n",
+    "\n",
+    "plt.legend(loc='best')\n"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.5.4"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}